Congruency, Expectations And Consumer Behavior In Digital Environments 365828420X, 9783658284206, 9783658284213

Table of contents :
Acknowledgments: A letter to myself.......Page 6
Contents......Page 8
Tables......Page 10
Figures......Page 12
List of Abbreviations......Page 13
1.1 Empirical Relevance of Digital Environments......Page 15
1.2 Central Domains and Theoretical Foundation in the Context of Digital Environments......Page 16
1.3 Research Gaps in Central Domains of Digital Environments......Page 19
2.1 Focus of the Essays......Page 27
2.2 Essay 1 - Does Mental Construal Influence the Perception of Incongruent Advertisement? The Role of Psychological Distance in Ad Processing......Page 28
2.3 Essay 2 - Vividness of Product Images in Online Stores: The Role of Delivery Time......Page 30
2.4 Essay 3 - Is It Human? The Role of Anthropomorphism as a Driver for the Successful Acceptance of Digital Voice Assistants......Page 32
2.5 Essay 4 - Online Complaints in the Eye of the Beholder: Optimal Handling of Consumer Complaints on the Internet......Page 35
2.6 Essay 5 - Differences and Similarities in Motivation for Offline and Online eSports Event Consumption......Page 37
2.7 Essay 6 - The Need for a Community: The Impact of Social Features on Video Game Success......Page 39
2.8 Overview of Essays and Related Research Characteristics......Page 42
3.1.1 Introduction......Page 43
3.1.2 Literature Review and Hypotheses Development......Page 45
3.1.3.1 Study 1......Page 50
3.1.3.2 Study 2......Page 54
3.1.3.3 Study 3......Page 60
3.1.4 General Discussion......Page 68
3.1.5 Appendix......Page 72
3.2.1 Introduction......Page 74
3.2.2 Literature Review and Hypotheses......Page 77
3.2.3 Empirical Studies......Page 84
3.2.3.1 Study 1......Page 85
3.2.3.2 Study 2......Page 92
3.2.4 Discussion and Conclusions......Page 102
3.2.5 Appendix......Page 107
3.3.1 Introduction......Page 109
3.3.2 Conceptual framework and hypotheses development......Page 111
3.3.3 Impact of the anthropomorphism on digital voice assistants......Page 112
3.3.4 Technological acceptance drivers for digital voice assistants......Page 115
3.3.5 Method......Page 118
3.3.6 Results......Page 120
3.3.7 General Discussion and Implications......Page 123
3.3.8 Appendix......Page 126
3.4.1 Introduction......Page 129
3.4.2 Framework......Page 130
3.4.3 Method......Page 132
3.4.4 Results and Implications......Page 133
3.4.5 Conclusion......Page 138
3.4.6 Appendix......Page 140
3.5.1 Introduction......Page 142
3.5.2 Literature review and hypothesis development......Page 144
3.5.3 The empirical study......Page 154
3.5.4 Conclusion......Page 158
3.5.5 Appendix......Page 160
3.6.1 Introduction......Page 162
3.6.2 Conceptual Framework and Hypotheses Development......Page 164
3.6.3 Impact of the product features......Page 167
3.6.4 Impact of social features......Page 169
3.6.5 Method, Research Design and Sample......Page 171
3.6.6 Results......Page 173
3.6.7 Discussion and Conclusion......Page 177
4.1 Core Results and Conclusions......Page 182
4.2 Research and Theoretical Implications......Page 188
4.3 Managerial Implications......Page 191
4.4 Directions for Future Research......Page 195
References......Page 200

Citation preview

Bernhard Swoboda • Thomas Foscht Hanna Schramm-Klein Hrsg.

Frederic Nimmermann

Congruency, Expectations and Consumer Behavior in Digital Environments

Handel und Internationales Marketing Retailing and International Marketing

Series Editors Bernhard Swoboda, Trier, Germany Thomas Foscht, Graz, Austria Hanna Schramm-Klein, Siegen, Germany

Die Schriftenreihe fördert die Themengebiete Handel und Internationales Marketing. Diese charakterisieren – jedes für sich, aber auch in inhaltlicher Kombination – die Forschungsschwerpunkte der Herausgeber. Beide Themengebiete werden grundsätz­ lich breit aufgefasst; die Reihe bietet sowohl Dissertationen und Habilitationen als auch Tagungs- und Sammelbänden mit unterschiedlicher inhaltlicher und metho­ discher Ausrichtung ein Forum. Die inhaltliche Breite ist sowohl im Sinne eines konsumentenorientierten Marketings wie auch einer marktorientierten Unternehmensführung zu verstehen. Neben den Arbeiten, die von den Herausgebern für die Schriftenreihe vorgeschlagen werden, steht die Reihe auch externen wissenschaftlichen Arbeiten offen. Diese können bei den Herausgebern eingereicht und nach einer positiven Begutachtung publiziert werden. The book series focuses on the fields of Retailing and International Marketing. These two areas represent the research fields of the editors—each of them as a single research area, but also in combination. Both of these research areas are widely understood. Consequently, the series provides a platform for the publi­ cation of doctoral theses and habilitations, conference proceedings and edited books, as well as related methodological issues that encompass the focus of the series. The series is broad in the sense that it covers academic works in the area of consumer-oriented marketing as well as the area of marketoriented management. In addition to academic works recommended by the editors, the book series also welcomes other academic contributions. These may be submitted to the editors and will be published in the book series after a positive assessment.

More information about this series at http://www.springer.com/series/12697

Frederic Nimmermann

Congruency, Expectations and Consumer Behavior in Digital Environments

Frederic Nimmermann University of Siegen Siegen, Germany Dissertation, University of Siegen, 2019

ISSN 2626-3327 ISSN 2626-3335  (electronic) Handel und Internationales Marketing Retailing and International Marketing ISBN 978-3-658-28420-6 ISBN 978-3-658-28421-3  (eBook) https://doi.org/10.1007/978-3-658-28421-3 Springer Gabler © Springer Fachmedien Wiesbaden GmbH, part of Springer Nature 2020 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer Gabler imprint is published by the registered company Springer Fachmedien Wiesbaden GmbH part of Springer Nature. The registered company address is: Abraham-Lincoln-Str. 46, 65189 Wiesbaden, Germany

Acknowledgments: A letter to myself. Dear Future Me, what a surprise that you are reading again your dissertation. I hope you will always remember the people around you, who contributed during this time to your work and life. No matter if colleagues, superiors, friends, family or especially your better half: they all shaped memories regarding a more than important stage of your life and contributed considerably to the success of your doctorate. So please always remember... ... your supervisor, Prof. Dr. Hanna Schramm-Klein, who always gave you important impulses for your scientific work and a corresponding freedom to shape your research as you wanted, who gave you constructive criticism, but who also knew when she had to encourage you, who opened new perspectives for your professional life and who helped you to discover a little more of this world, ... the advisory committee members of your dissertation, Prof. Dr. Marcus Schweitzer, who supported the final preparation of this dissertation with meaningful ideas in the run-up to completion and, above all, showed a considerable degree of flexibility in terms of his time and also Prof. Dr. Arndt Werner, who was available at short notice for your disputation and took over the chairmanship of the associated committee, ... Carmen Richter, who has always been the good soul at the chair and has supported you in your tasks in everyday university madness, who was your no. 1 proofreader and who was always available for a good “cup of coffee-chat”, ... your colleagues (and partially co-authors) at that time, Florian Neus, Robér Rollin, Florentine Frentz, Theresia Mennekes, Anne Fota, Katja Wagner, Tobias Röding and especially your scientific mentors PD Dr. Sascha Steinmann, Dr. Gerhard Wagner and Dr. Gunnar Mau, for the exciting lunchtime roundtables, the joint adventures at conferences and the support in professional as well as private life questions, ... to your parents, Fabienne and Wilhelm Nimmermann, who always kept your back free for your studies, put their own wishes aside to support you and found the right words even in difficult moments, ... and especially your fiancée, Jana Freund, for her continuous support in the last years: her patience, especially in difficult times, her encouragement, her willingness for discussions and feedback on your research and also her endless support in the years before that led to the dissertation in the first place. Be grateful for every moment with her. Right now, you are.

VI

Acknowledgments: A letter to myself.

Your connection to people in your life and their contribution to your work and private life always only make sense in retrospect. Hence, future me and everybody else reading these lines, stay open for new things, behave in a way that creates new possibilities and as someone famous once said: “stay hungry, stay foolish”.

Siegen, Frederic Nimmermann

Contents Acknowledgments: A letter to myself. ................................................................................................. V Tables .................................................................................................................................................... IX Figures .................................................................................................................................................. XI List of Abbreviations ........................................................................................................................ XIII 1 Introduction.................................................................................................................................... 1 1.1 Empirical Relevance of Digital Environments ....................................................................... 1 1.2 Central Domains and Theoretical Foundation in the Context of Digital Environments .......................................................................................................................... 2 1.3 Research Gaps in Central Domains of Digital Environments................................................. 5 2 Structure and Content of the Essays .......................................................................................... 13 2.1 Focus of the Essays ............................................................................................................... 13 2.2 Essay 1 - Does Mental Construal Influence the Perception of Incongruent Advertisement? The Role of Psychological Distance in Ad Processing .............................. 14 2.3 Essay 2 - Vividness of Product Images in Online Stores: The Role of Delivery Time ........ 16 2.4 Essay 3 - Is It Human? The Role of Anthropomorphism as a Driver for the Successful Acceptance of Digital Voice Assistants ................................................................................ 18 2.5 Essay 4 - Online Complaints in the Eye of the Beholder: Optimal Handling of Consumer Complaints on the Internet .................................................................................. 21 2.6 Essay 5 - Differences and Similarities in Motivation for Offline and Online eSports Event Consumption ............................................................................................................... 23 2.7 Essay 6 - The Need for a Community: The Impact of Social Features on Video Game Success .................................................................................................................................. 25 2.8 Overview of Essays and Related Research Characteristics .................................................. 28 3 Essays ............................................................................................................................................ 29 3.1 Does Mental Construal Influence the Perception of Incongruent Advertisement? The Role of Psychological Distance in Ad Processing ......................................................... 29 3.1.1 Introduction ................................................................................................................. 29 3.1.2 Literature Review and Hypotheses Development ........................................................ 31 3.1.3 Experiments ................................................................................................................. 36 3.1.3.1 Study 1 ........................................................................................................... 36 3.1.3.2 Study 2 ........................................................................................................... 40 3.1.3.3 Study 3 ........................................................................................................... 46 3.1.4 General Discussion ..................................................................................................... 54 3.1.5 Appendix ...................................................................................................................... 58 3.2 Vividness of Product Images in Online Stores: The Role of Delivery Time ........................ 60 3.2.1 Introduction ................................................................................................................. 60 3.2.2 Literature Review and Hypotheses .............................................................................. 63 3.2.3 Empirical Studies......................................................................................................... 70 3.2.3.1 Study 1 ........................................................................................................... 71 3.2.3.2 Study 2 ........................................................................................................... 78 3.2.4 Discussion and Conclusions ........................................................................................ 88 3.2.5 Appendix ...................................................................................................................... 93

VIII

4

Contents 3.3 Is It Human? The Role of Anthropomorphism as a Driver for the Successful Acceptance of Digital Voice Assistants ................................................................................ 95 3.3.1 Introduction ................................................................................................................. 95 3.3.2 Conceptual framework and hypotheses development.................................................. 97 3.3.3 Impact of the anthropomorphism on digital voice assistants ...................................... 98 3.3.4 Technological acceptance drivers for digital voice assistants .................................. 101 3.3.5 Method ....................................................................................................................... 104 3.3.6 Results........................................................................................................................ 106 3.3.7 General Discussion and Implications ....................................................................... 109 3.3.8 Appendix .................................................................................................................... 112 3.4 Online Complaints in the Eye of the Beholder: Optimal Handling of Consumer Complaints on the Internet .................................................................................................. 115 3.4.1 Introduction ............................................................................................................... 115 3.4.2 Framework ................................................................................................................ 116 3.4.3 Method ....................................................................................................................... 118 3.4.4 Results and Implications............................................................................................ 119 3.4.5 Conclusion ................................................................................................................. 124 3.4.6 Appendix .................................................................................................................... 126 3.5 Differences and Similarities in Motivation for Offline and Online eSports Event Consumption ....................................................................................................................... 128 3.5.1 Introduction ............................................................................................................... 128 3.5.2 Literature review and hypothesis development ......................................................... 130 3.5.3 The empirical study ................................................................................................... 140 3.5.4 Conclusion ................................................................................................................. 144 3.5.5 Appendix .................................................................................................................... 146 3.6 The Need for a Community: The Impact of Social Features on Video Game Success ...... 148 3.6.1 Introduction ............................................................................................................... 148 3.6.2 Conceptual Framework and Hypotheses Development ............................................ 150 3.6.3 Impact of the product features ................................................................................... 153 3.6.4 Impact of social features ........................................................................................... 155 3.6.5 Method, Research Design and Sample ...................................................................... 157 3.6.6 Results........................................................................................................................ 159 3.6.7 Discussion and Conclusion ....................................................................................... 163 General Discussion and Conclusion ......................................................................................... 169 4.1 Core Results and Conclusions............................................................................................. 169 4.2 Research and Theoretical Implications ............................................................................... 175 4.3 Managerial Implications ..................................................................................................... 178 4.4 Directions for Future Research ........................................................................................... 182

References ........................................................................................................................................... 187

Tables Table 1.3-1. General Research Goals ....................................................................................................... 6 Table 2.1-1. Essay Overview.................................................................................................................. 13 Table 2.8-1. Summary of Essays and Research Characteristics ............................................................. 28 Table 3.1-1. ANOVA Results (DV = Congruence) ............................................................................... 39 Table 3.1-2. MANOVA Results – Pre-Experiment Brand and Advertising congruence ....................... 43 Table 3.1-3. MANOVA Results (DV = Congruence) ............................................................................ 44 Table 3.1-4. ANOVA Results (DV = Advertising Congruence post experiment) ................................. 50 Table 3.1-5. ANOVA Results (DV = Implementation, Attitude and Credibility) ................................. 51 Table 3.1-6. ANOVA Results (DV = Interest in more Information) ..................................................... 52 Table 3.1-7. Overview Constructs Essay 1 ............................................................................................ 58 Table 3.2-1. Results of Hypotheses Testing ANOVA (DV = Mental Imagery) .................................... 74 Table 3.2-2. Results of Hypotheses Testing ANOVA (DV = Attitude towards the Presented Product) ................................................................. 75 Table 3.2-3. Results of Hypotheses Testing ANOVA (DV = Purchase Intention) ................................ 75 Table 3.2-4. Results of Hypotheses Testing ANOVA (DV = Construal Level) .................................... 76 Table 3.2-5. Results of Hypotheses Testing ANOVA (DV = Attitude towards the Presented Product) ................................................................. 77 Table 3.2-6. Results of Hypotheses Testing ANOVA (DV = Purchase Intention) ................................ 77 Table 3.2-7. Results of Hypotheses Testing ANOVA (DV = Attitude towards product -utilitarian and hedonic) ................................................. 83 Table 3.2-8. Results of Hypotheses Testing ANOVA (DV = Purchase Intention) ................................ 84 Table 3.2-9. Results of Hypotheses Testing ANOVA (DV = Ownership and Consumption Imagery). 85 Table 3.2-10. Overview Constructs Essay 2 .......................................................................................... 93 Table 3.3-1. Results of PLS-SEM ........................................................................................................ 106 Table 3.3-2. Overview Constructs Essay 3 .......................................................................................... 112 Table 3.3-3. Correlations, Squared Correlations and Average Variance Extracted ............................. 114 Table 3.4-1. Types of organizational response to a public complaint .................................................. 116 Table 3.4-2. Main effects of organizational response on observer’s behavior ..................................... 119 Table 3.4-3. Mean values of behavioral variables across different types of organizational response . 120 Table 3.4-4. Mediating Effects ............................................................................................................. 123 Table 3.4-5. Overview Constructs Essay 4 .......................................................................................... 126

X

Tables

Table 3.5-1. Hypothesis testing ............................................................................................................ 142 Table 3.5-2. Overview Constructs Essay 5 .......................................................................................... 146 Table 3.6-1. Variable operationalization .............................................................................................. 158 Table 3.6-2. Report of the results ......................................................................................................... 161

Figures Figure 1.2-1. General Conceptual Framework ......................................................................................... 5 Figure 3.1-1. Conceptual Model ............................................................................................................. 31 Figure 3.1-2. Inner-condition Differences of Advertising Congruence Perception, CG = congruence, CL = construal level ........................................................................... 40 Figure 3.1-3. Schematic flow of the video prime in study 2 .................................................................. 41 Figure 3.1-4. Delta-Congruence-Perception, CG = congruence, PD = psychological distance ............. 45 Figure 3.1-5. Schematic flow of the video-prime in experiment 3 ........................................................ 47 Figure 3.2-1. Conceptual Model ............................................................................................................. 63 Figure 3.2-2. Manipulation of Vividness in Study 1 .............................................................................. 71 Figure 3.2-3. Product page with coffee machine in low vividness and high vividness conditions ........ 79 Figure 3.2-4. Moderating effects of mental imagery ability on the interaction of delivery time and vividness for purchase intention ................................................................................ 87 Figure 3.3-1. Conceptual Framework ..................................................................................................... 98 Figure 3.4-1. Organizational responses to public complaints and their impact on behavior of an observer of public complaint handling .................................................................. 117 Figure 3.5-1. Conceptual Model ........................................................................................................... 130 Figure 3.6-1. Conceptual Framework ................................................................................................... 152

List of Abbreviations Ad ............................................. advertisement AI .............................................. artificial intelligence ANCOVA ................................. analysis of covariance ANOVA .................................... Analysis of variance AP ............................................. average playtime app ............................................ application AVE .......................................... average variance extracted B ............................................... regression coefficient BIF ............................................ Behavior Identification Form CASA ....................................... Computers Are Social Actors CG............................................. congruency CL ............................................. construal level CLT........................................... construal level theory DLC .......................................... downloadable content DV ............................................ dependent variable ECA .......................................... embodied conversational agents ECT........................................... expectation-confirmation theory ESL ........................................... Electronic Sports League eSports ...................................... electronic sports EU LCS .................................... European League of Legends Championship Series F ................................................ F-statistic GPS ........................................... Global Positioning System H ............................................... hypothesis HCI ........................................... Human Computer Interaction IPA ............................................ intelligent personal assistants LLCI ......................................... lower level of confidence interval M............................................... mean MANCOVA ............................. multivariate analysis of covariance MANOVA ................................ multivariate analysis of variance MIA .......................................... mental imagery ability MS ............................................ mean squares MSSC ....................................... motivation scale for sports consumption N ............................................... number of sample size

XIV

List of Abbreviations

n ................................................ number of sub-sample size η² ............................................... Eta-squared n.s.............................................. not significant (Ne)WOM................................. (negative electronic) Word of Mouth NFI ............................................ normed fit index NOV ......................................... number of owners of the video game p ................................................ p-value p. ............................................... page PAD .......................................... Pleasure-Arousal-Dominance PD ............................................. psychological distance PLS ........................................... partial least squares PR ............................................. price R ............................................... Pearson’s r (correlation coefficient) R2 .............................................. R-squared (coefficient of determination) RQ............................................. research question SD ............................................ standard deviation SEM .......................................... structural equation model sig. ............................................ significance level SPSS ......................................... Statistical Package for the Social Sciences SRMR ....................................... standardized root mean square residual t ................................................. t-statistic ULCI ......................................... upper level of confidence interval UTAUT2................................... unified theory of acceptance and use of technology 2 VIF ............................................ variance inflation factor VVIQ ........................................ Vividness of Visual Imagery Questionnaire α ................................................ Cronbach’s alpha β ................................................ beta (standardized coefficient)

1 Introduction 1.1 Empirical Relevance of Digital Environments Digital environments are “virtual or cyber-generated environments that [be] can accessed or created through the use of one or more digital devices such as a computer, tablet, or a mobile phone” (Dowell, 2019, p. 128). Environments such as the internet have become one of the most important marketplaces in the world (Kemp, 2018) and offer challenges due to developments in digital marketing (Leeflang et al., 2014). Hence, a general objective for research is, to get a more profound view on how digital environments affect consumers (Stephen, 2016). Consequently, because of the importance of digital environments for consumers and businesses, this thesis strives to deepen knowledge in this area. In general, digital environments such as streaming services, video games, or the web itself with its digital shopping stores and chat and discussion communities have affected the daily lives of consumers tremendously. More than four billion consumers use the web on a daily basis with an estimated yearly growth of 7% globally (Kemp, 2018). Hence, these environments affect how consumers form their overall daily life (e.g., what consumers do) and on which devices they rely on, on a day-to-day basis, e.g., smartphones or laptops. More than 5.135 billion consumers use mobile, web-enabled devices such as smartphones and spend around six hours per day using the web and related services, such as streaming, online shopping, or gaming (Kemp, 2018). Forecasts predict that more than 80% of all internet users in the United States will access digital video content on platforms such as YouTube or streaming providers such as Netflix daily in 2019 (Statista, 2019a; 2019c). Phenomena such as electronic sports (eSports) support this rise of streaming services. With more than 385 million viewers worldwide, eSports are one of the most important streaming service topics (Freeman and Wohn, 2017; Hamari and Sjöblom, 2017). In particular, an average viewing session in streaming services is about 40 minutes long, which has increased by 50% from 2018 (Aslam, 2019). Another example of digital environments is the use of video games; twothirds of all US citizens regularly play video games (Entertainment Software Association, 2016). In general, 2.4 billion consumers worldwide will play video games in 2019, due to the use of mobile devices (McDonald, 2019). As an additional example, more than 45% of all internet users use the web regularly for online shopping, resulting in more than 1.7 billion consumers purchasing fashion, electronics, food, travel, digital services, and toys on the web (Kemp, 2018).

© Springer Fachmedien Wiesbaden GmbH, part of Springer Nature 2020 F. Nimmermann, Congruency, Expectations and Consumer Behavior in Digital Environments, Handel und Internationales Marketing Retailing and International Marketing, https://doi.org/10.1007/978-3-658-28421-3_1

2

1 Introduction

However, the rise of these digital environments not only affects consumers, but also how a business aligns its activities, e.g., its advertising efforts, products, and services, based on this usage. This example of e-commerce has enormous potential for retail companies in terms of participation. E-commerce sales worldwide are predicted to raise more than 3,453 billion US dollars in 2019 (Statista, 2019b). With the huge impact of video games on leisure activities, it is not surprising that video games contribute to almost 53 billion US dollars in spending in terms of video game sales (Kemp, 2018). Based on the usage of streaming services or social media, which have a worldwide user base of more than 3.196 billion users (Chaffey, 2019), these digital environments offer outstanding advertising opportunities. Hence, brands and companies are keen to spend more than 37 billion US dollars in video and social media advertising in 2022, representing a growth of 130% in only five years (Foye, 2018). 1.2 Central Domains and Theoretical Foundation in the Context of Digital Environments In summary, digital environments already have a huge impact on consumers and businesses alike and considering the growth of each subareas (e.g., streaming consumption), one can assume that these environments will continue to have a considerable influence and achieve ubiquity in the years to come (Hootsuite, 2019). Generally, in recent years new advances in internet technology can be summarized by the transformation into Web 2.0 and even further to Web 3.0 due to the emergence of semantic web technologies and their impact on how consumers interact with digital environments (Berners-Lee et al., 2001; Garrigos Simon et al., 2012). This digital revolution within digital environments creates tremendous challenges for business and research alike, such as consequences of new digital channels and media (Leeflang et al., 2014). Garrigos‐Simon et al. (2012) stress that progress in these technologies and the increasing expansion and use of digital environments are leading to remarkable shifts regarding effective business activities. Hence, the overall assumption is that digital environments are evolving and that these transformations not only impact business but also affect consumers’ “attitudes, beliefs, and practices” (Barassi and Treré, 2012, p. 1270). Research needs to examine consumers' media practices to provide more accurate analyses of the diverse use of digital environments and its impact on social contexts and lived experiences (Barassi and Treré, 2012). Thus, these assumptions lead to a need for a holistic approach for research in digital marketing, including manifold subareas and relevant topics related to information acquisition, processing and decision aids in digital environments, and shoppers’ behavior and advertising (Kannan and Li, 2016; Yuping, 2019). To do so, this thesis investigates both dimensions of digital environments: “environment–integral” (digital environments that influence behavior from within) or

1.2 Central Domains and Theoretical Foundation in the Context of Digital Environments

3

“environment–incidental” (digital environments that influence behavior in unrelated environments). These are of particular relevance (Stephen, 2016, p. 18) when addressing digital environments using a more holistic approach. Based on this reasoning, digital environments still offer a broad range of general research objectives that address key issues in central domains (Keller, 2014; Mela, 2018; Roy et al., 2016; Sheoran et al., 2018; Stephen, 2016). The thesis sets itself the goal to focus on central domains, which are marked by their ubiquity in consumers’ daily lives and which form primary consumer behavior in digital environments (or are, at least, directly related to that behavior) (Bawden and Robinson, 2008; Kemp, 2018). For instance, based on previous reasoning and empirical relevance, e-commerce (i.e., digital shopping) is a key element of digital environment usage (Kemp, 2018). However, consumers shopping online are exposed to related advertising, they seek new information prior to a purchase or interact with platforms using digital assistants (Stephen, 2016). Another example is the use of new media forms and topics, such as using streaming services for digital gaming broadcasts (Hamari and Sjöblom, 2017); these, yet again, affect how consumers are exposed to different forms of advertising or information processing. Hence, these central domains are related to each other and can play a particularly important role in each of the areas under review. It seems feasible to obtain a more profound understanding of consumer behavior and processing in these central domains. Consequently, this thesis seeks to contribute to the following general research objectives in key digital domains: n n n n n

n

Advertising: investigating the advertising processing of consumers in digital environments (e.g., Bishop et al., 2015); E-Commerce: investigating shoppers’ information processing in the case of online shopping (e.g., Yoo and Kim, 2014); (Mobile) Conversational Agents: investigating the interaction with the digital voice assistants of those environments (e.g., Seeger et al., 2017); Complaint-Management: enhancing the understanding of complaint management for information-seeking web consumers (e.g., Zhang et al., 2010); New Media Forms: broadening the knowledge regarding the differences in motivation for electronic sports consumption between offline (event on-site) and digital environments (event stream) (e.g., Hamari and Sjöblom, 2017); and Digital Gaming: the impact of social elements of an online community in digital environments as success drivers for video games (e.g., Butler et al., 2014).

In recent years, research has already addressed the different phenomena of these environments to enhance knowledge on consumer behavior in these contexts and has called for further investigation (Lamberton and Stephen, 2016). This is, likewise, in line

4

1 Introduction

with general calls for research in marketing (e.g., Keller, 2014; Mela, 2018). It is against this backdrop that this thesis seeks to understand how the developments and specific phenomena in digital environments affect consumer behavior and processing. Hence, it aims to addresses these research objectives by taking relevant specific phenomenon in each subarea into account. Focusing on these research objectives, this thesis builds on two fundamental assumptions, which are related to each other (Maille and Fleck, 2011): consumer expectations (e.g., Oliver, 1977) and congruency effects (e.g., Osgood and Tannenbaum, 1955). In general, these theoretical approaches are of particular importance in marketing-related research (e.g., Maille and Fleck, 2011; Spreng and Olshavsky, 1993) and offer explanations for the effects under review in this thesis. Very briefly, consumers create expectations prior to an action such as buying or searching for information (Oliver, 1977). If these expectations are fulfilled, resulting in positive disconfirmation, consumers have a sense of satisfaction or, in cases of negative disconfirmation, dissatisfaction (Oliver, 1980). Research offers multiple approaches for explaining these comparison standards (Spreng and Olshavsky, 1993), such as the value–percept disparity, in which the comparison is facilitated in a “[…] response triggered by a cognitive-evaluative process in which the perceptions of (or beliefs about) an object, action, or condition are compared to one's values (or needs, wants, desires)” (Westbrook and Reilly, 1983, p. 258). Hence, consumers are exposed to their own expectations and behave accordingly (Spreng and Olshavsky, 1993). Regarding these comparisons, congruence effects are in line with these assumptions by taking consistency models into account (Tannenbaum, 1967). In general, congruence refers to the matching of or similarity between objects, i.e., objects are consistent with each other or at least have a structural correspondence based on knowledge and related expectations (Maille and Fleck, 2011). In particular, literature stresses the importance of schemata (i.e., prior knowledge structure) because of their impact on the evaluation of congruence due to consumer expectations (Mandler, 1982; Stayman et al., 1992). The positive effects of congruence can be explained based on the assumption that individuals value the harmony of the given combination in general, or the specific elements of combined objects (e.g., context and ads) (Osgood and Tannenbaum, 1955). Hence, in both cases, the literature proposes that the convergence of assumptions or perceptions lead to more positive processing and evaluation; these, in turn, lead to favorable outcomes such as the positive behavior of consumers. In summary, this thesis investigates phenomena in digital environments that affect consumer behavior as well as attitudinal responses; however, these phenomena are based on cognitive processing and evaluations, which are based on the general assumptions of expectations and congruency effects. These fundamental assumptions are depicted in Figure 1.2-1.

1.3 Research Gaps in Central Domains of Digital Environments

5

Digital Environments Cogni&ve Processing and Evalua&on

Expecta&ons

Congruency

Phenomena in Digital Environments

Consumer Behavior and A7tudinal Responses

Central Domains Adver&sing E-Commerce (Mobile ) Conversa&onal Agents Complaint-Management New Media Forms Digital Gaming

Addi&onal Influen&al Factors



Figure 1.2-1. General Conceptual Framework

While this theoretical framework proposes the basic concept of this thesis, each essay is accompanied by a well-grounded research framework related to its specific context and builds on its corresponding research questions. More precisely, these assumptions play a particularly important role in this thesis, because congruency and expectation effects are based on general assumptions for consumer behavior, either directly, e.g., congruency effects in advertising in essay 1 or expectations regarding service recovery in essay 4, or indirectly, such as expectations due to different motivations, as in essay 5. However, specific research questions addressing the previously mentioned central domains shape the foundation of each essay, which will be presented in the following chapter. 1.3 Research Gaps in Central Domains of Digital Environments Based on previous reasoning, this thesis focuses on central domains and phenomena within digital environments. The main and subordinate purpose is, therefore, to obtain a more profound understanding about the consumers in these environments. In general, the thesis follows three main and subordinate research goals (see Table 1.3-1): first, this thesis strives for a deeper understanding of consumers’ cognitive processing in digital environments because previous research indicates peculiarities regarding the selection and processing of information (e.g., Ismagilova et al., 2019; Metzger and Flanagin, 2013; Yuping, 2019). Second, this thesis investigates the role of expectations in digital environments, which lead to decision making in choice and behavior based on previous research that stresses the general importance of those aspirations to explain consumer

6

1 Introduction

behavior regarding digital environments or technology (e.g., Burke, 2002; Butler et al., 2014; Hennig-Thurau and Walsh, 2004; Venkatesh et al., 2012). Third and final, since previous literature stresses the importance of dissimilarities between digital and analog environments (e.g., Stephen, 2016; Yoo and Kim, 2014; Zhang and Byon, 2017), this thesis investigates how these particularities shape the behavior of consumers for specific subjects. Table 1.3-1. General Research Goals Topics of Interest A profounder understanding of…

… consumers’ cognitive processing in digital environments.

Essay 1

Essay 2

P

P

… expectations affecting choice and behavior in digital environments. … differences of digital environments to analogues ones.

Essay 3

Essay 5

Essay 6

P P

P

Essay 4

P

P P

These research goals are in line with the proposed relationships in the general conceptual framework (see Figure 1.2-1): in essence, the thesis discusses how cognitive processing in various sub-areas of digital environments is influenced by expectations and congruence effects. However, it strives to do so by examining important research gaps within the six central domains. These domains, consolidated within digital environments, each have specific research questions, so that an overview can be successively presented in the following chapter. Congruency effects are an important issue in the first domain, which is related to advertising (Dahlen et al., 2008). Research in this domain stresses the importance of contextual congruency between advertisements and the actual context in which it is shown (Yuping, 2019). In particular, studies in this domain show effects of ad-context congruency in different digital environments such as the web or in video games. Thematic congruency generates positive reactions to an advertisement (Moore et al., 2005), positive attitudes towards the advertised brand (Choi and Rifon, 2002), and stronger purchase intentions (Jeong and King, 2010). Research has also shown positive effects of incongruency, for instance, with regard to memory effects; however, this seems heavily dependent on incongruency operationalization (e.g., Dahlen et al., 2008; Halkias and Kokkinaki, 2014; Moore et al., 2005). However, due to the rise of digital environments, and, thus, advertising in different contexts (e.g., video streams or mobile games with different topics, genres, and content) it is difficult for advertisers to guarantee a thematic congruence between their products or services and the actual context (Li and Lo, 2014). Consequently, it is not surprising that literature calls for additional research to achieve maximum advertising effectiveness (Yuping, 2019) and

1.3 Research Gaps in Central Domains of Digital Environments

7

in particular, a deeper understanding of interaction between digital context and advertisements (Mela, 2018). The previously mentioned effects can be explained by schema theory, which postulates that incoming information that does not match pre-existing knowledge, and, thus, cannot be processed in relation to preexisting knowledge, leads to a perception of incongruency (Mandler, 1982). In particular, a high level of incongruence and the need for strong cognitive adjustments to fit incoming information in previous knowledge, leads to frustrated consumers and negative evaluations (d Astous and Bitz, 1995). However, research shows that cognitive flexibility helps find possible connections within given knowledge and, thus, can enhance the resolution of incongruence (Jhang et al., 2012; Meyers-Levy and Tybout, 1989). Hence, an important research objective is to find ways to enhance this resolution. Surprisingly, the impact of psychological distance as anchors (e.g., in time) for digital advertisements has not yet been analyzed. Based on construal level theory, it is psychological distance that leads to concrete or abstract mental construal in consumers (Trope and Liberman, 2010). Transferring the basic assumptions of schema (incongruity) theory (Mandler, 1982) to the study, this thesis proposes that an abstract mental construal leads to an enhanced resolution of incongruency in a digital context, whereas a concrete mental construal is more effective in processing congruent information. Thus, in the domain of advertising, this thesis investigates how psychological distances might be used to enhance the resolution of digital incongruent advertisements. Hence, the first research question arises: n

How do psychological distances affect the congruency perception and processing of advertisements in digital environments?

Effects of vividness and mental imagery are of particularly importance for retailing in digital environments such as the internet (Jiang and Benbasat, 2007; Yoo and Kim, 2014), which is the second domain investigated in this thesis. The lack of haptic and gustatory information, e.g., being able to a touch a given product in online environments can lead to purchase cancellations due to reduced confidence in decision making (Peck and Childers, 2003). However, research stresses a higher purchase probability for products that are presented with higher sensory richness (e.g., Steinmann et al., 2014; Van Der Heide et al., 2013). Yet, product pictures represent a substantial source of sensory information in an online store (Yoo and Kim, 2014). Thus, online retailers try to tackle this issue by offering vivid product presentations (e.g., a product picture) and higher sensory richness, which, in turn, lead to more positive outcomes (Coyle and Thorson, 2001). This can be explained by the effects of mental imagery, i.e., a mental event involving the visualization of a concept or relationship (Lutz and Lutz, 1978). Here, mental imagery reflects the process by which a sensory experience is represented

8

1 Introduction

in the consumers’ working memory (MacInnis and Price, 1987). Shoppers experience the gratification that arises from consuming alternative products presented more vividly, which, in turn, leads to more favorable outcomes for retailers (Coyle and Thorson, 2001). However, while most research about online retailing has focused on individual elements of web environments (Chung et al., 2018), such as interactivity (e.g., Xu and Sundar, 2012), sensory stimuli in general (e.g., Parsons and Conroy, 2006) and vividness in particular (e.g., Steinmann et al., 2014), surprisingly, one key element has been ignored by taking mental imagery into account: delivery time as moderator. Previous research concerning delivery time mostly interprets it as a barrier to e-commerce (e.g., Chen et al., 2010; Li et al., 2014). However, this thesis investigates delivery time as a form of psychological distance. Previous research has shown that psychological distance might shape how information heuristics affect consumer behavior (Braga et al., 2015). Hence, by taking the construal level theory into account, the thesis proposes that temporal distance might shape the way that abstract or concrete information are preferred in the context of online retailing. In particular, it proposes that abstract information is preferred due to a high mental construal evoked by a high (temporal) psychological distance and vice versa. Thus, the second research question arises: n

How do psychological distances shape preferences in processing concrete or abstract product information on online stores?

As well as the internet, digital environments also affect consumer behavior in the analog world (Lamberton and Stephen, 2016). Hence, taking a more holistic approach to digital environments into account, another research question deals with digital voice assistants and the impact of human-like elements on the behavioral intentions of consumers. In general, striving for the most realistic human-like design of conversional agents is seen as a desirable goal in research (Seeger et al., 2017). According to the media equation theory (Reeves and Nass, 1996), consumers apply social norms based on interactions with humans when interacting with various media such as computers and robots. So far, research has shown that a well-balanced anthropomorphic phenomenon, known as the "persona effect", is believed to have promoted the credibility of non-human actors and how they positively affect consumers' attitudes toward the system (Lester et al., 1997). These assumptions are in line with the Computers Are Social Actors (CASA) paradigm, which proposes that consumers apply identical social heuristics used for human interactions with computers (Nass and Moon, 2000) and, consequently, reflects the human–human trust perspective (Nass et al., 1994). The CASA paradigm represents an established theoretical basis for research with an interest in understanding how to enhance the trustworthiness of digital agents (Riedl et al., 2014; Seeger et al., 2017). This leads to consumers being more responsive toward computers than they would be

1.3 Research Gaps in Central Domains of Digital Environments

9

to a person (Nass et al., 1994) and pay attention to presenting one’s self positively (Sproull et al., 1996). However, in studies regarding interactions with visual agents, consumers also showed nervousness toward overly intense observations by the agent (Brahnam and De Angeli, 2012). Research has also shown negative effects of anthropomorphism in case of robots (MacDorman and Ishiguro, 2006). In particular, these results are coherent with the socalled Uncanny Valley Paradox (Mori, 1970), which describes the effect in which a human-like design leads to an improvement in users’ perceptions but only to a certain degree and only until the similarity is so strong that it is perceived uncanny (Mori, 1970; Mori et al., 2012). Consequently, the design of the robot is inconsistent because, on the one hand, it is congruent with a real human, but on the other, it is already too advanced to be clearly classified as robotic. This creates a situation in which the robot cannot (immediately) be assigned to a category, so that the Uncanny Valley effect arises (Mori, 1970). Hence, a research question arises about how beneficial human-like elements are based on these contradicting views about the accepted use of voice assistants. This question contributes to technology acceptance models such as unified theory of acceptance and use of technology (UTAUT2) (Venkatesh et al., 2012) because humanlike elements are not yet part of these considerations, and, thus, allow a more holistic view of this technology. So here, the third research question arises: n

How does a perceived congruency due to human-like elements of digital voice assistants and actual users affect intentions to use such technology?

Digital environments offer a broad range of information that consumers consider prior to a purchase decision (Hennig-Thurau and Walsh, 2004); this is the fourth area of specification in this thesis. In general, research on complaint handling has developed strategies regarding its management, i.e., how a company should respond to a complaint that is related to dimensions of redress, timeliness, facilitation, apology, credibility, and attentiveness (Davidow, 2003). Research stresses the importance of those established dimensions and considers their impact on consumer behavior (e.g., Einwiller and Steilen, 2015; Estelami, 2000; Sparks and McColl-Kennedy, 2001). Research shows that an effective use of complaint strategies and successful complaint handling cannot only recover the satisfaction of the complaint (Homburg and Fürst, 2007), but also lead to an increase in the satisfaction level of the complainant regarding the company, which goes beyond a level at which no service failure happened, and thus, leads to the so-called service recovery paradox (McCollough et al., 2000). Research in this context has mostly focused on the actual complainant. The recent years of digitalization have had an impact on general methods of communication (Deighton

10

1 Introduction

and Kornfeld, 2009), which means that these complaints are also publicly available on the internet (Hennig-Thurau et al., 2010) and are almost infinitely accessible (Duan et al., 2008). Hence, research should distinguish in the context of these strategies between the actual complainant and those consumers who seek information prior to a purchase (Hennig-Thurau and Walsh, 2004). Previous theoretical concepts, such as expectation confirmation theory (e.g., Oliver, 1980; Oliver and DeSarbo, 1988), justice theories (e.g., Bies and Moag, 1987; Blodgett et al., 1997; Cook and Hegtvedt, 1983), and cognitive dissonance theory (Festinger, 1957) offer potential explanations for the underlying effects of the processing of companies’ answers to complainant research. Specifically, a congruency between the consumers’ expectations of how the company should answer and how the actual service recovery (i.e., the response to a complaint) shapes attitudinal and behavioral outcomes (Homburg and Fürst, 2005). While these theoretical assumptions might also hold true for the silent observer, the weighting and processing might differ due to different motivations (Hennig-Thurau and Walsh, 2004; Hennig-Thurau et al., 2004) and reference points (Smith et al., 1999). Thus, the fourth research question addresses the expectations of consumers and how a company should answer to a service failure of an unknown complainant and the corresponding processing and related behavior. Hence, the fourth research question arises: n

How do “silent observers” of a public consumer complaint process the answer of a company by considering established service recovery strategies?

Recent years have extended traditional means for consumers’ digital consumption, such as streaming services (Bründl et al., 2017). The fifth domain of this thesis deals with the opportunity for event consumption in digital environments and the related underlying motivation of consumers. Previous research has investigated leisure motivation (Beard and Ragheb, 1983), which has been extensively drawn out to different event forms (Li and Petrick, 2005), such as festivals (e.g., Nicholson and Pearce, 2001), universal expositions (Lee, Lee and Wicks, 2004), concerts (e.g., Kulczynski et al., 2016), or sports consumption in general (Trail and James, 2001) and literature stressed the differences between these event forms. However, technological developments and advances in digital environments led to complete new event forms (Zhang and Byon, 2017) and related industries such as eSports (Freeman and Wohn, 2017; Hamari and Sjöblom, 2017). However, while there is established research regarding motivation for event sport consumption (Kirkup and Sutherland, 2015), there is a specific research gap in dealing with eSports. Following the uses and gratification theory (Katz, Blumler, et al., 1973), consumers decide on consumption based on their interests (e.g., context) and needs (e.g., escape from reality or entertainment) offered by the media (Hamari et al., 2018). The decision regarding a medium, thus, depends on user expectations and the satisfaction of

1.3 Research Gaps in Central Domains of Digital Environments

11

the needs of the media offer; therefore, developing congruency between expectations and fulfillment (Severin and Tankard, 2013). However, while digital consumption forms such as streaming services are popular for many sports forms, eSports are rooted in digital environments, whereas classic sports (e.g., soccer or basketball) originate from an analog world of consumption (Hamari and Sjöblom, 2017; Zhang and Byon, 2017). Consequently, one can observe a shift from pure digital consumption to alternatives, such as a digital consumption in an offline environment, which might fulfill different needs. Even though research has already investigated differences in offline and online event consumption, indicating differences between these environments, it is mostly based on specific forms of sports (e.g., Hu et al., 2017; Seo and Green, 2008; Zhang and Byon, 2017), research in the context of eSports is scarce. Hamari and Sjöblom (2017) introduced an adapted motivation scale for consumption, especially for eSports. Building on that idea, Pizzo et al. (2017) analyzed the differences between sports and eSports consumption in relation to a given sports theme. However, research has not yet distinguished between the two forms of consumption based on the environment (i.e., digital vs. analog) for eSports. Hence, a more profound understanding of those environments contributes to building a foundation for further efforts on theory and model building of future research (Li and Petrick, 2005). Thus, the fifth research question arises: n

How does event participation motivation as spectators differ between analog and digital environments for eSports?

Finally, digital environments offer opportunities for leisure activities, in which gaming has become one of the most important possibilities and the final domain under research in this thesis. Despite the importance of video games, success drivers of video games have not yet been sufficiently empirically investigated. In general, research proposes that the success of a video game depends on the high number of users of a specific video game console, e.g., Xbox or PlayStation (Gretz, 2010). Surely, a high number of users of a specific video game console has a positive effect on the sales volume of video games due to the user base (Gallaugher and Wang, 2002). However, the quality offered in video games seems to be an even more important factor (Anderson et al., 2014; Lee, 2013). In fact, research interprets video games as hedonic products (e.g., Lin, 2010; Turel et al., 2010), so quality can only be assessed by a consumer after a purchase is process. Thus, video games as hedonic products portray economic risks for new users (Voss et al., 2003); consumers rely on third-party information to reduce risk (Hennig-Thurau and Walsh, 2004). Previous research shows that consumers rely on the reputation of the developer and reviews from journalists (Cox, 2013), which affect consumers’ expectations and video game sales (Binken and Stremersch, 2009). The game genre, the

12

1 Introduction

actual pricing, and the promotion effect shape the actual success of a video game in a similar way to that of movies (Cox, 2013; Hennig-Thurau et al., 2012; Marchand, 2016). Nevertheless, the role of newer features such as social interactions with other gamers while playing video games have been ignored as potential success drivers so far. However, the way video games are consumed by users through social media, online forums, and multiplayer features has changed dramatically within the last ten years (Marchand and Hennig-Thurau, 2013). Social interaction, knowledge-sharing, and lively discussions are essential factors of any online community (Butler, 2001; Wasko and Faraj, 2005) and the video game community thrives on the same principles. In this context, network effect theories (e.g., Beck, 2006) suggest that the value for consumers of a membership in a network is “positively affected when another user joins and enlarges the network” (Katz and Shapiro, 1994, p. 94). Consequently, the benefits of consuming a product such a video game depend positively on the total number of owners who purchased it, because it enhances the product due to its incentive for the provision of complementary products or services (Church et al., 2008). Hence, through new online multiplayer features, in-app purchases, and communication channels, users have been able to update video games, add new features, and share their passion with people from different countries and continents. Video games are no longer just a product that a user buys for a fixed price and enjoys playing on their own or with known friends. The perceived utility of these multiplayer and communication features goes beyond the traditional gaming experience as it enhances interpersonal relationships (Ledbetter and Kuznekoff, 2011). Such relationships satisfy a fundamental need for relatedness through interpersonal social interactions within the virtual world (Downie et al., 2008). However, these social interactions have not been empirically analyzed in this context. Transferring the basic assumption of Oliver’s (1977) expectation–confirmation theory to this setting, the thesis investigates the impact of social interactions on video game success (e.g., number of owners) to set ground for a more holistic approach in video game success research. Thus, another research question arises: n

How do social interactions shape the success of a video game in comparison to classical video game elements?

Based on previous reasoning, this thesis strives to answer relevant research questions in central domains of digital environments. Investigating these specific questions and phenomena in each area adds meaningful knowledge to previous research and sets new foundations for further research. These research questions are in line with practical issues in digital worlds, so that both academia and practitioners profit from the results and implications of the thesis.

2 Structure and Content of the Essays 2.1 Focus of the Essays The primary goal of this work is to enhance understanding of the central domains of consumers’ cognitive processing, their behavior in digital environments, and, moreover, how expectations and congruence explain and shape the central consumer related outcomes as proposed in the general conceptual framework. All six essays (see Table 2.1-1) focus on specific topics in digital environments, by dealing with relevant questions for practice, and, in particular, by taking important marketing academiarelated research questions into account. Table 2.1-1. Essay Overview Essay Title

Subarea of Digital Environments

Does Mental Construal Influence the Perception of Incongruent Advertisement? The Role of Psychological Distance in Ad Processing

Digital Advertising

Vividness of Product Images in Online Stores: The Role of Delivery Time

Online Retailing

Is It Human? The Role of Anthropomorphism as a Driver for the Digital Voice Assistants Successful Acceptance of Digital Voice Assistants Online Complaints in the Eye of the Beholder: Optimal Handling of Consumer Complaints on the Internet

Online Complaint Management

Differences and Similarities in Motivation for Offline and Online eSports Event Consumption

Digital Sports Consumption

The Need for a Community: The Impact of Social Features on Video Game Success

Online Communities

This chapter contains an overview over the central essays included in this thesis by presenting extended abstracts of the content. The first essay focuses on the role of psychological distances based on the construal level theory (Trope and Liberman, 2010) for processing advertisements in streaming services and mobile games. In particular, the essay focuses on how context incongruent brands profit from abstract mental construal, and, thus, a broader schemata of consumers. The second essay addresses the impact of delivery time as a psychological distance on the effects of a congruence between concrete or abstract mental construal, and more or fewer vivid product images in online stores. Essay three investigates the role of anthropomorphism for the acceptance of digital voice assistants, i.e., how the congruence between the assistant’s voice and © Springer Fachmedien Wiesbaden GmbH, part of Springer Nature 2020 F. Nimmermann, Congruency, Expectations and Consumer Behavior in Digital Environments, Handel und Internationales Marketing Retailing and International Marketing, https://doi.org/10.1007/978-3-658-28421-3_2

14

2 Structure and Content of the Essays

human-like characteristics impacts the intention to use a given assistant. The fourth essay focusses on the effects of different complaint-handling strategies of businesses in online environments, while, at the same time, shifting the perception of these strategies to the “silent observer” and their expectations in comparison with well-established research regarding the complainant. The fifth essay addresses the difference in motivation for online and offline event consumption for eSport events and which expectations towards these events are shaped (e.g., possibilities of social interaction). The last and sixth essay targets the influence of social features in video games for their success and how this might affect consumers’ expectations for the introduction of new video games. In the following subchapters, the purpose, the central research questions, and the methodology of each essay are summarized. The main results and contributions are also presented. 2.2 Essay 1 - Does Mental Construal Influence the Perception of Incongruent Advertisement? The Role of Psychological Distance in Ad Processing The purpose of the first essay is to introduce psychological distances as a novel influential factor for perception advertisements in digital environments, and, thus, broadening the knowledge for research and practice for advertising processing. Increasing digitalization in recent years has had a large impact on consumers’ consumption of services in digital environments. eSports have earned, with more than 385 million viewers worldwide, the role of one of the most important streaming service topics (Freeman and Wohn, 2017; Hamari and Sjöblom, 2017). Generally, digital gaming is one of the most popular forms of entertainment (Herrewijn and Poels, 2015) and this industry has become one of the largest entertainment industries (McDonald, 2017). This highlights the potential of advertising in these environments, and, therefore, their importance for both marketing research and practice. In this context, the thematic congruence of streaming or game content (e.g., automobile blog or racing game) and advertisement (e.g., automobile advertising) is of particular importance and is regarded as one of the most important criteria for the selection of an advertising medium (Dahlen et al., 2008). Based on schemata, research stresses the importance of a contextual advertising congruence because it leads to positive effects on consumer behavior and evaluations of advertising (e.g., Bishop et al., 2015; Moore et al., 2005). In contrast, incongruity usually leads to negative behavior and evaluations (Halkias and Kokkinaki, 2014). This can be explained due to the ease of categorization of the advertising input and context to previous knowledge, (e.g., a jeans advertisement on a fashion website), and, thus, an enhanced and more fluent advertising processing (e.g., Halkias and Kokkinaki, 2014; Mandler, 1982; Verberckmoes et al., 2016). In contrast, incongruency needs a higher level of cognitive processing to find a link between context and

2.2 Essay 1 - Impact of Mental Construal on Advertising Processing

15

advertising at all, which leads to more challenging categorization and, thus, more negative evaluations (Houston et al., 1987). Due to the heterogeneity of the content of each of these services (e.g., YouTube) or the implementation of dynamic advertising systems in mobile games, it is a challenge for advertisers to display in a congruent context. Surprisingly, there is a research gap in the literature regarding the influence of schema incongruence and how to support the resolution of incongruence among consumers to counteract this challenge. Hence, based on construal level theory (Trope and Liberman, 2010), this essay postulates a relationship between mental abstraction levels due to invoked psychological distances from the advertisement and context congruence evaluation. Here, in conjunction with a high construal level and associated with a more abstract mental level, consumers should more easily find the "link" between a context-incongruent brand and the environment, so is more likely that a higher level of congruence is perceived. Conversely, in the case of a context-congruent brand, a concrete level of abstraction should, additionally, reinforce the obvious schema links. To close the research gap and verify the assumptions regarding the effects of psychological distances, three experiments (2x3 betweensubject design) were carried out in the context of advertising in two streaming services and in a racing game for mobile devices. In all three experiments, respondents had to evaluate a given advertisement in a congruent (vs. medium congruent vs. incongruent) context. In the first experiment (soccer stream, N = 228), the construal level was operationalized as the disposition of the participants, whereas in the follow-up experiments the psychological distance (temporal "long vs. short" and geographical "far vs. near") in the context of an event announcement (experiment 2, streaming service, N = 214) and announcement of a game expansion (experiment 3, mobile racing game, N = 221) was artificially stimulated. To analyze and interpret the data of the three experiments, analysis of variance (ANOVA), related contrast tests, and regression analyses were conducted. The preliminary study (experiment 1) confirmed the general impact of different level of mental construal on the congruence evaluation. Results of the two consecutive followup experiments show the successful altering of the mental level of abstraction due to a low vs. high psychological distance from the advertisement and, therefore, different evaluations of the three brands (low, medium, and high congruence within the context) depicted in the experiments. Lower and, thus, more concrete mental abstraction levels promoted the perception of advertising of context-congruent brands, whereas a high level of mental construal was beneficial in case of an incongruent brand (e.g., advertising jeans in a racing game). These effects are not only observable in case of congruency processing, but also for the overall processing of the advertisement, e.g., the

16

2 Structure and Content of the Essays

advertisement’s credibility. Thus, matching a given level of congruency with an analog level of mental construal, due to a specific level of psychological distance, leads to a more positive advertising process and positively affects behavioral outcomes.One main contribution of the first essay is its expansion of marketing knowledge by introducing psychological distances as a novel influential factor for the processing of congruency, especially in the case of advertising in digital environments such as streaming services and mobile games. Practitioners benefit from new advertising strategies; brands desiring to advertise in an incongruent context (e.g., due to a favorable target group) or in a less controllable environment (e.g., YouTube videos) should advertise in a psychologically distanced and, thus, more abstract context, to promote the ease of the incongruent advertisement’s cognitive processing and, therefore, positively support its effects. 2.3 Essay 2 - Vividness of Product Images in Online Stores: The Role of Delivery Time The purpose of the second essay is to investigate the role of delivery time as psychological distance for consumers’ processing of product presentations in online stores, broadening understanding of online shopping by considering delivery time not only as a barrier for e-commerce, but as an important moderator for attitudinal and behavioral responses of shoppers. E-commerce is one of the most important growth sectors worldwide: global retail ecommerce sales are expected to reach the mark $ 4.878 billion by 2021 (eMarketer, 2018). Thus, retailers see an extraordinary incentive to participate in this sector. However, online retailers have to optimize their stores, e.g., product presentations, in the best possible way to attract online shoppers. Here, product images of a given product are particularly important, as, in contrast to brick and mortar stores, the sensory richness is heavily reduced. Products on online stores cannot be touched or smelled, which is a major challenge for online retailing (Yoo and Kim, 2014). However, research has shown that a more pronounced sensory richness (e.g., a higher degree of haptics) leads to a higher level of purchase confidence, which results in more positive behavioral outcomes (e.g., Müller, 2013; Van Der Heide et al., 2013). One explanation offered by research is that a higher level of sensory richness, which is the representational richness of a mediated environment, leads to a higher level of vividness for a given product (Steuer, 1992). Here, the mental imagery reflects the process by which a sensory experience is represented in consumers’ working memory (MacInnis and Price, 1987). Based on this reasoning, shoppers experience a higher level of consumption gratification arising from products presented more vividly in contrast to a less vivid presentation, which is assumed to lead to more favorable behavioral outcomes and thus, concluding that in an

2.3 Essay 2 - Vividness of Product Images in Online Stores: The Role of Delivery Time

17

online environment a higher level of vividness of product images leads to more positive consumer responses (Coyle and Thorson, 2001). This assumption highlights a research gap, because research surprisingly ignores one central characteristic of online stores as a factor that not only produces transactions costs but leads to a psychological distance: the delivery time. Based on construal level theory (Trope et al., 2007), research suggests a connection between temporal distances (e.g., delivery time) and the mental construal of consumers (Trope and Liberman, 2010) and moreover, that consumers tend to overcome psychological distances by creating analog mental representations (Lynch and Zauberman, 2007). In addition, research indicates relationships between these mental abstraction levels and the heuristics used by consumers (e.g., more abstract information) (Braga et al., 2015). However, it is unknown how this affects the preferences for different levels of product image vividness by considering created mental representations. Thus, this essay investigates the impact of different mental abstraction levels (i.e., based on the delivery time) on the preference of different product images on online stores and their relationship to behavioral and attitudinal outcomes of shoppers. To close the research gap and verify assumptions regarding effects of temporal distances on product image processing, this essay contains two consecutive experiments. Both experiments followed a 2x2 between-subject design with the delivery time of the experimental factors (high vs. low) and vividness of the product images (high vs. low). In both experiments, participants had to evaluate a product on an artificial online store and answer questions regarding their attitudinal and behavioral responses, correspondingly. Here, the products were depicted in different combinations of vividness and delivery times. In the first experiment (N = 256), respondents had to evaluate a chocolate bar as a hedonic product, which was wrapped (low vividness condition) and unwrapped (high vividness condition). The first experiment investigated whether the proposed effects held true and how mental imagery was affected by a congruence through an abstract (vs. concrete) mental construal and a more abstract (vs. concrete) product image due to their vividness. The second experiment (N = 365) extended this by adding a more utilitarian product (i.e., a printer) in addition to a hedonic product (i.e., a coffee machine) for robustness tests of the results regarding different product categories. The second study expanded on the findings regarding the mediating effects of mental imagery and shed light on potential boundaries in which the effects disappeared. To analyze and interpret the data of the two experiments, analysis of (co-) variance (ANCOVA) and related contrast tests were used. Potential moderating effects were investigated using regression analyses.

18

2 Structure and Content of the Essays

The results of the first study showed that a higher delivery time leads to a higher construal level, which leads to the conclusion that delivery time should be interpreted as psychological distance. While results do not show an expected main effect of the vividness (i.e., higher vividness is more favorable), results suggest that matching the delivery time with a corresponding vividness level of the product image might be suitable in case of low delivery time, and, thus, more concrete mental construal, because consumers prefer a higher level of vividness in product images. In contrast, in the case of a long delivery time and, therefore, processing with a higher level of mental construal, a product image that is less vivid and more abstract leads to a more favorable outcome. Results of the second study confirm the previous ones and testing both for a more hedonic and utilitarian product indicates an independency of the effects regarding the product category. In addition, results show that ownership imagery as a subdimension of mental imagery is affected by the interplay of delivery time and level of vividness; however, surprisingly, consumption imagery is only affected by the direct effects of vividness. Finally, results also suggest that the proposed effects only occur if the prospective shopper has the ability for analog imagination. In cases of low mental imagery ability, results indicate that the effects of the interplay of delivery time and vividness disappear. The main contribution of the second essay is that it shows that the delivery time in online stores should not been seen only as a barrier during the shopping process but as an important moderator that affects the processing during shopping and, thus, attitudinal and behavioral responses. Previous research indicates the effects of psychological distances in offline contexts (e.g., perception of assortment), which opens further research questions in digital shopping environments, i.e., which effects of delivery time and other psychological distances go beyond affecting perceptions of vividness. For online retailers, this study shows that it seems beneficial to implement a dynamic product image rotation based on delivery time, so that more concrete or abstract product images are given to prospective shoppers. 2.4 Essay 3 - Is It Human? The Role of Anthropomorphism as a Driver for the Successful Acceptance of Digital Voice Assistants The third essay focuses on the impact of the human-like characteristics of digital voice assistants in comparison to classical drivers of technology acceptance on the intention to use those smart assistants. Digital voice assistants such as Apple’s Siri or Amazon’s Alexa are revolutionizing access to web content (e.g., shopping or weather information) and the use of smart technology (e.g., smart home devices). These assistants show tremendous growth; in the first three quarters of 2017, more than 17 million smart speakers were delivered

2.4 Essay 3 - Anthropomorphism and Digital Voice Assistants

19

worldwide and another 16 million during the holiday season (Gibbs, 2018). Experts estimate that by 2020, half of online searches worldwide will be made using the voice (Olson, 2016), which also indicates a shift in consumers’ overall use of technology. In general, intelligent personal assistants are applications that use inputs such as the user's voice or contextual information (e.g., GPS position) to assist in (verbally) answering natural language questions, as well as making recommendations and performing actions (Hauswald et al., 2015). Thus, assistants such as Siri use certain skills such as determining actions, solving problems, drawing inferences, and returning with a verbal corresponding solution or action for the user (Hayes-Roth, 1995). Here, a recent study shows that vocal interaction can actually trigger emotions (Horstmann et al., 2018). Thus, dealing with voice assistants not only indicates an interaction with technology per se, but an interaction with a human-like receiver and sender and, therefore, a duality of technology. Anthropomorphism, hence, seems of particularly importance; it is defined as the tendency to attribute the actual or perceived behavior of non-human actors, human characteristics, motivations, intentions, or emotions (Epley et al., 2007). Literature regarding, e.g., embodied conversational agents such as robots, indicates that anthropomorphism usually plays a key role in influencing the behavior of consumers interacting with these agents (Becker et al., 2007; Fridin and Belokopytov, 2014). Goudey and Bonnin (2016) show that a partially anthropomorphic appearance improves acceptance by consumers with practical experience of analogous technology. Waytz et al. show (2014) that anthropomorphism increases the overall trust in using the given technology. Hence, there are consequences of adding human characteristics to machines. Research does not yet offer studies that address the previously mentioned duality and there are no empirical insights into anthropomorphism concerning digital conversational agents and the assistants’ voice, or how it influences the intention to use a given assistant in comparison to classic technology drivers. Past research is mostly related to conversational agents dealing with technological aspects of the software (e.g., Bellegarda, 2013; Harris, 2005). Previous research refers to the effect of personification or the integration of emotions in the design of conversational agents, but without considering the users’ acceptance (e.g., Callejas et al., 2011). The use of natural language in studies concerning digital conversational agents and the anthropomorphism of the voice is usually of minor importance. Thus, identifying a research gap, the fifth essay addresses three central research questions regarding the duality of voice assistants: first, which role does anthropomorphism take for the acceptance of digital voice assistants? Second, which drivers of anthropomorphism (i.e., perceived sociability, animacy and human-like fit) are the most important to affect the likability of the

20

2 Structure and Content of the Essays

assistant? Third, how does anthropomorphism compare with regard to the importance of the intention to use a digital voice assistant in comparison to classical technology drivers? To address these research questions, a quantitative online survey was conducted (N = 283). First, participants had to answer questions regarding regular technology acceptance drivers such as habit or performance expectancy based on the unified theory of acceptance and use of technology. Second, participants had to evaluate potential anthropomorphism drivers such as perceived sociability, animacy, and a human-like fit considering the voice of the voice assistant. In addition, participants had to evaluate how these drivers affect the likeability of the assistant. In order to ensure a preexisting experience of the participants with voice assistants, e.g., on a smartphone, participants were asked how familiar they are with voice assistants, how often they use them or watch someone using this technology, and with which assistants they are dealing (e.g., Siri, Alexa, Google Assistant, Cortana, Bixby, or others) on a regular basis. To analyze and interpret the data of the survey, structural equation modeling, based on partial least squares approach, was used. First of all, results show that most drivers of the classical technology acceptance approach (e.g., performance expectancy or habit) also hold true for digital voice assistants. As expected, results show that the higher the expected performance of the voice assistant and the higher the hedonic motivation of the consumer, the higher the intention to use a given assistant. However, the previous habit also substantially affect behavioral intentions. Four of the classical drivers (i.e., social influence, effort expectancy, facilitating conditions, and price value) show no significant impact on the intention to use a given assistant. Second, results also show a significant impact of all three predicted anthropomorphism drivers (i.e., perceived sociability, animacy and a perceived human-like fit) on the likeability of a voice assistant. This significantly influences the behavioral intention and, thus, confirms the importance of anthropomorphism for the acceptance of digital voice assistants. The central contribution of the third essay is it shows that when studying voice assistant acceptance using a more holistic approach regarding technology acceptance, drivers for this kind of technology go beyond the classical drivers of the unified theory of acceptance and use of technology and, therefore, a "conservative" view of technology is not sufficient. For future research, this essay offers a broad range of implications. For instance, human characteristics might be interpreted and valued differently depending on cultural background (House, 2004). Hence, multicultural research should be sought to ensure comparability and obtain a more complete view on the impact of anthropomorphism elements based on different cultural backgrounds. Future research could further investigate additional factors that are not covered in the study, such as the

2.5 Essay 4 - Online Complaints in the Eye of the Beholder

21

importance of being extensible by other devices in smart home technology. Besides, the essay also offers a broad range of implications for management, such as showing that it is not sufficient to only focus on the services of a voice assistant, but also to optimize the verbal interaction with the consumer. Implementing humor to conversations, which is a human trait, might enhance the overall experience with a voice assistant. 2.5 Essay 4 - Online Complaints in the Eye of the Beholder: Optimal Handling of Consumer Complaints on the Internet The objective of the fourth essay is to investigate different complaint-handling strategies in a digital environment (e.g., addressing a consumer complaint on a rating platform) by focusing on the “silent observer,” who seeks additional information regarding products or services. Dissatisfied consumers represent a potential hazard to a company and its business operations (Cheng et al., 2006). However, literature shows that effective complaint management can, not only limit these potential threats (Homburg and Fürst, 2007) but also lead successful problem solutions to increase the complainant’s satisfaction level with the company. This can even go beyond a level in which there is no service failure, thus, leading to the so-called service recovery paradox (McCollough et al., 2000). Strategies regarding complaint management, i.e., how a company should answer to a complaint, are well investigated in literature (Davidow, 2003) and research shows their importance considering the impact on consumer behavior (e.g., Collie et al., 2000; Conlon and Murray, 1996; Estelami, 2000; Sparks and McColl-Kennedy, 2001), although these strategies focus only on the actual complainant. The recent years of digitalization has had an impact on general modes of communication (Deighton and Kornfeld, 2009), so that these complaints are expressed and made public on the internet (Hennig-Thurau et al., 2010). Dissatisfied customers may post their complaint online to warn others against buying a particular product or service (Hennig-Thurau et al., 2004). This form of negative electronic word-of-mouth (NeWOM) can be accessed anywhere, at any time, via a broad internet audience (Cheung et al., 2008) and it is almost infinitely accessible (Duan et al., 2008). Above all, this audience includes prospective customers who search for information about a product or brand with the intention of assessing its quality and forming their attitudes and preferences prior to making a purchase decision (Hennig-Thurau and Walsh, 2004). Thus, this form of complaint is even more of a hazard for businesses because it affects potential costumers (Chevalier and Mayzlin, 2006). If one takes the strategies to answer to these complaints as the status quo, it is unclear how prospective customers, who search for information on the web, are affected by the

22

2 Structure and Content of the Essays

public answers of a given company and, thus, lead to a research gap. Based on previous reasoning, an organizational response to a public complaint on the internet might go beyond the complaint handling and satisfaction recovery of the complainant, because it might entail additional effects on the attitudes, intentions, and behavior of prospective customers who simply observe the conversation between an organization and a customer in a complaint-handling setting. To address this research gap, an explorative unifactorial between-subject experiment (N = 233) with nine treatments in total has been conducted. The participants of the study were asked to recall the last time they bought a product from an online store and to transfer themselves mentally to a depicted scenario. Here, respondents were told that they found the product was cheaper in an unknown online store; however, on a website on which other customers discuss this specific shop, they found a (negative) customer experience report, which was depicted in the questionnaire afterwards. This report was followed by a response of the online shop formulated in accordance with one of the eight types of organizational responses contained in the research framework (Davidow, 2003). In addition, a positive review was added as a control group. After being exposed to the scenario, respondents had to evaluate the online store based on the customer’s review and the corresponding response by the company. To analyze and interpret the data of the experiment, analysis of (co-)variance (ANCOVA) and corresponding contrast tests were used. The results of the experiment demonstrate that the answer of a company to a public complaint affects the prospective customer who seeks additional information prior to a purchase. Both attitudinal and behavioral intentions vary based on the response of the company. Here, the essay offers three central results regarding the use of the common strategies by taking the silent observer into account. First, offering an apology and redress is not only one of the most powerful strategies for the complainant (Davidow, 2003), it is also for the prospective customer. Here, the positive impact of the answer regarding consumer-related variables even surpasses the control group with the absence of a problem, so an effect similar to the service-recovery paradox occurs. Second, in contrast to how it affects the complainant, the strategy of a request for a private contact (i.e., by phone or mail), leads the observer of the complaint handling to an outcome, which is even worse than the passive handling of the given complaint, i.e., no answer at all (Homburg and Fürst, 2007). Even so, this strategy is characterized by a high level of attentiveness and, thus, is powerful to address the complainant, it does not offer additional information to the silent observer because of the absence of an answer (i.e., more than a request for contact). Hence, in this case, results show a gap between a favorable strategy in case of the complainant in comparison to the prospective customer. Third, if a redress is not possible and cannot be offered, the worst and best strategies do

2.6 Essay 5 - Motivation for Offline and Online eSports Event Consumption

23

not differ significantly in terms of their effect on behavior, so companies have a variety of answer strategies to select from. The central contribution of the third essay is that it shows that strategies regarding an answer to a complaint that are based on wellestablished research also affect public digital environments and prospective customers. Thus, the status quo should be further analyzed by considering the aspect of public complaints, which are accessible by almost everyone. This essay offers implications for companies on how to answer complaints by not only respecting the complainant but also by considering stimulating consumers who seek additional information prior to making a purchase decision. 2.6 Essay 5 - Differences and Similarities in Motivation for Offline and Online eSports Event Consumption The fifth essay focuses on the dimensions of consumers’ motivation to participate at an eSports event in an offline or digital online environment based on the motivation scale for sports consumption (MSSC) and, hereby, identify the central differences and similarities that might lead to the consumption of one these event forms. Increasing digitalization in recent years has had a large impact on consumers’ consumption phenomena, in, or arising from, digital environments. For instance, eSports are already a key phenomenon, which can be roughly defined as “a form of sports where the primary aspects of the sport are facilitated by electronic systems” (Hamari and Sjöblom, 2017). In particular, eSports have earned the role of one of the most important streaming topics with more than 385 million viewers worldwide (Freeman and Wohn, 2017; Hamari and Sjöblom, 2017). However, eSports offline events, often hosted in big arenas, allow thousands of eSports fans who are willing to leave the purely digital environment of streams to consume eSports content in a completely new setting (Hallmann and Giel, 2017). The content, i.e., following teams competing in a digital environment, is identical in both the stream and offline event. However, many of the surrounding factors vary (e.g., the perception of cheering spectators) and might change the overall experience (Zhang and Byon, 2017). Although research has focused on aspects of general eSports consumptions, it does not deal with different forms of eSports consumption. In this context, digital consumption forms, such as streaming services, are popular for many sports forms. However, eSports roots in digital environments, whereas classical sports (e.g., soccer) originate from an analog world of consumption (Hamari and Sjöblom, 2017; Zhang and Byon, 2017). Hence, one might observe a shift in the case of eSports from pure digital consumption to analogous opportunities of consumption (e.g., event on-site). Literature has offered approaches to classify eSports in relation to other

24

2 Structure and Content of the Essays

digital phenomena, as well as traditional sports and related differences (Hallmann and Giel, 2017; Macey and Hamari, 2017). Research has also worked out motivation dimensions of general eSports consumption: For instance, Hamari and Sjöblom (2017) established a motivation scale that is especially tailored to eSports. Building on that idea, Pizzo et al. (2017)analyzed the differences between sports and eSports consumption for a specific sports form such as soccer. However, while research indicates the differences between offline and online consumption for regular sports and the motivation to follow an event (e.g., Seo and Green, 2008; Zhang and Byon, 2017), surprisingly, a comparison of the two previously described forms of eSports consumption has not yet been made. In this context, the uses and gratification theory (Katz, Blumler, et al., 1973) proposes that consumers decide on the basis of interests (e.g., formats) and needs (e.g., information seeking) from which media form is consumed (Hamari et al., 2018). Consequently, the actual decision for a medium depends on user expectations and fulfilment, leading to a satisfaction of the needs by the media offer (Severin and Tankard, 2013). Hence, based on potential different needs (Katz, Blumler, et al., 1973), the question arises as to what motivates consumers to attend an event on-site or to follow the given stream online and choose a specific media form. This is of particular interest, because traditional sports have made the transition to being additionally consumable in digital environments (e.g., streams), whereas, in comparison, eSports originate from these digital environments and are now present in offline environments. To address this research question, a study (N = 636) at an EU LCS Event in Berlin was conducted in early 2018. The event was one the biggest competitions for League of Legends, one of the most important eSports games (Cocke, 2018). Riot Games, the organizer of the event, offered exclusive live coverage of the event through lolsports, YouTube, and Twitch.tv and, thus, the online availability of this event via streams is an alternative to attending the event on-site. Therefore, the survey distinguished between recipients who visited the event on-site (n = 155) and consumers who followed the event via one of the streams (n = 482). Here, participants had to answer questions based on the MSSC for eSports (Hamari and Sjöblom, 2017). Participants gave insights into their attitudinal responses to the event and potential behavioral responses for upcoming events. To analyze and interpret the experiment data, students’ t-tests were used. The results showed a difference between motivation for sports consumption for those who visited the event on site and those who consumed the matches via a digital streaming service. Social elements seem to foster the decision to participate in an offline environment, e.g., to be part of a cheering crowd, the dimensions of learning new skills, gaining knowledge, and a higher level of achievement. However, elements that can be observed more in more detail on screens at home via stream led to an online

2.7 Essay 6 - The Need for a Community: The Impact of Social Features on Video Game Success

25

consumption decision. Surprisingly, both forms of event consumption had no impact on the attitude towards the event. However, the overall satisfaction was more pronounced in case of offline consumption. Finally, the results showed that, in both cases, further online consumption is very likely, although an intention for offline consumption was far more pronounced in the case of those who were already at the on-site event. The results showed that a specific motivation was underlying the current consumption decision form, which might also affect future consumption decisions in the case of eSports. The central contribution of the fourth essay is to show that, while the actual content of an eSports event is similar in both cases of consumption (i.e., offline vs. online), the surroundings seem to determine which consumption form one might decide upon based on his motivation for sports consumption. Hence, further research should consider that motivation for eSports consumption does not only shape the general intention to watch a given match or event (Hamari and Sjöblom, 2017), but also the form that is chosen in which to consume. Therefore, research should distinguish between offline consumption and a participation in digital environments by respecting each influential factor that surrounds these environments and that might have an impact on consumer behavior. Practice should promote the elements that support the specific motivation of online and offline participants and, thus, optimize their consumption experiences in the chosen environment. 2.7 Essay 6 - The Need for a Community: The Impact of Social Features on Video Game Success The sixth essay focusses on the impact of social features in comparison to more common core elements (e.g., price) of video games on their success based on market data from one of the most important gaming platforms. Video games have become tremendously popular: the industry is one of the largest entertainment industries globally, with an increase from $101 billion in 2016 to an expected $129 billion in 2020 (McDonald, 2017). Two thirds of all US citizens play video games on a regular basis (Entertainment Software Association, 2016). Herrewijn and Poels (2015) regard digital gaming as one of the most popular forms of entertainment. Despite their popularity, the success drivers of video games have not yet been sufficiently empirically investigated, although this an important issue for developers. The production costs of video games also increased tremendously due to new hardware technologies and state-of-the-art game mechanics, graphics, and sound effects. The production budget of “Grand Theft Auto V,” one of the most successful video games of

26

2 Structure and Content of the Essays

all time, was around $250 million (IMDB, 2016); thus, creating such high-quality games is linked to a considerable financial risk for developers. However, taking this risk can certainly pay off; video games that end up being successful can generate more than 300% of their original production costs (Clements and Ohashi, 2005). The previously mentioned video game “Grand Theft Auto V” generated about $800 million in revenue within the first 24 hours of its release (IGN, 2013). There is high uncertainty among developers regarding which type of video game or core elements should be considered (e.g., in-app purchases); thus far, success drivers of video games have not been entirely empirically identified. Research has shown that the success of a given video game depends on a high number of users of a specific video game console (Gretz, 2010) and on the reputation of the developer and reviews from journalists (Cox, 2013). Here, Binken and Stremersch (2009) showed a direct positive effect of professional reviews on the expectation and sales of video games. The game genre, actual pricing, and promotion effect, similar to movies, render the actual success of a video game (Cox, 2013; Hennig-Thurau et al., 2012; Marchand, 2016). However, surprisingly, the role of newer social features (e.g., multiplayer and social [surrounding] interactions with other users while playing the video game or in forums on given platforms such as Steam) have been ignored as potential success drivers, or have been only partially investigated for the success of other game features, e.g., multiplayer for in-game sales (Marchand, 2016). However, it is not only the core features that might impact the expectation and satisfaction towards a video game but also the community’s respective social features that have formed around a video game. The perceived utility of those social features goes beyond the traditional gaming experience as it enhances interpersonal relationships, which satisfy a fundamental need for relatedness through interpersonal social interactions within the virtual world (Downie et al., 2008; Ledbetter and Kuznekoff, 2011). These assumptions are in line with network effect theories (e.g., Beck, 2006), which propose that the value of membership for consumers in a network is positively influenced when new consumers join and expand it (Katz and Shapiro, 1994). Consequently, the benefits of consuming such a video game depend on the total number of consumers who purchase it (Church et al., 2008). Thus, video games might no longer be just a product that consumers buy for a fixed price and enjoy playing on their own. Social interaction, such as knowledge-sharing and lively discussions are essential factors of any online community and the video game community thrives on the same principles (Butler, 2001; Wasko and Faraj, 2005). Transferring the basic assumption of Oliver’s (1977) expectation–confirmation theory to this setting, the research question arises regarding how social features affect the success of a given video game and how these features affect the average playtime, which could positively affect revenue over time (e.g., due to in-game purchases) (Marchand, 2016).

2.7 Essay 6 - The Need for a Community: The Impact of Social Features on Video Game Success

27

In terms of a response to the research question, market data seems the most feasible method. although publishers and developers do not make sales records of video games publicly accessible. For this reason, most recent studies (Cox, 2013; Marchand, 2016) that have examined success drivers of video games have collected data from third-party websites such as vgchartz.com. Unfortunately, it is not possible to check the validity of such data. Thus, to address this research question, the analysis of market data based on 401 video games from the Steam platform has been conducted. Steam offers popular video games as well as features for communication and social interactions among its users. Therefore, data (e.g., number of positive votes, price or social features) from the popular platform Steam directly via steamdb was used (Steam, 2018). If social drivers (e.g., a multiplayer feature) affects the expectations of users and fulfils them, they should influence the satisfaction with a given videogame, which, in turn, should be verifiable by the predicted relationships in the essay through the market analysis of the number of owners and the average playtime as the dependent variables. To analyze and interpret the market data, structural equation modelling based on the partial least squares approach was used. Results of the analyses confirm the predicted impact of social features on video games success. Both for the number of owners of a video game and its average play time, the multiplayer feature had a positive impact. Positive votes (i.e., reviews) of other gamers on that platform affected the actual number of owners of a given videogame positively, whereas the average playtime was positively influenced by lively discussion threads. The results also indicate the impact of the more common core elements such as genre, developer reputation and additional (in-game) content from the developer on both success variables. Furthermore, results remain stable when controlling the price and days since release. The central contribution of this last essay is that it sheds light on the success drivers of video games by taking into consideration the market data from a given platform. For research, results of this essay suggest that video games are more than just a product bought by a user at a fixed price that they enjoy playing alone. Instead, they become more a social experience in which people from all over the world play together, communicate, and share knowledge. Thus, future research should investigate video game research-related content through the lens of social interactions. For practice, it is important to acknowledge that even video games with a specific single player experience might also benefit from social features, because both discussion threads (e.g., for additional free content) and positive votes of other gamers positively affect the success of such a game. Thus, with this in mind, even developers of single-player games should be keen on stimulating their user base on participation in social surroundings, positively affecting their published games.

28

2 Structure and Content of the Essays

2.8 Overview of Essays and Related Research Characteristics Table 2.8-1 provides an overview of the six essays and the corresponding research characteristics. For each essay, the research objective is summarized and information about the research design is provided. The sample size for each study as well as various methodologies used is illustrated. In summary, more than 2,400 respondents were acquired to participate in online surveys or experimental studies and more than 400 video games were investigated in the last essay to test causal relationships proposed in the underlying conceptual frameworks, to expand the knowledge for research and propose meaningful implications for practice in digital environments. Table 2.8-1. Summary of Essays and Research Characteristics Objective

Design

Sample Size

Methodology

Essay 1

Investigating the impact Study 1–3: Online of psychological distances on congruency Experiments processing and evaluation

Study 1: N = 228 Study 2: N = 214 Study 3: N = 221

(M)ANOVA, Planned Contrast Tests

Essay 2

Investigating the congruency of mental construal and visual stimulus on behavior

Study 1: N = 256 Study 2: N = 365

AN(C)OVA, Planned Contrast Tests

Essay 3

Investigate the impact of human-like characteristics of smart assistants on technology acceptance

Online Survey

N = 283

PLS-SEM

Essay 4

Investigate the impact of complaint handling strategies on prospective customers in an online environment

Online Experiment

N = 233

(M)AN(C)OVA Planned Contrast Tests

Essay 5

Identify differences and similarities in motivation for eSports consumption in offline vs. digital environments

Field Study and Online Survey

N = 636 offline n = 155 online n = 481

t-tests

Essay 6

Investigate the impact of social elements on video game success

Market Data

N = 401 (video games)

PLS-SEM

Study 1–2: Online Experiments

3 Essays 3.1 Does Mental Construal Influence the Perception of Incongruent Advertisement? The Role of Psychological Distance in Ad Processing 3.1.1 Introduction Digitalization has had a large impact on consumers’ consumption of digital services, the use of digital ecosystems in general and numerous phenomena within these environments. For instance, “eSports” (electronic sports) is an important phenomenon within the rise of streaming services. With more than 385 million viewers worldwide, eSports is one of the most important streaming service topics (Freeman and Wohn, 2017; Hamari and Sjöblom, 2017). A second phenomenon is mobile games. Herrewijn and Poels (2015) regard digital gaming as one of the most popular forms of entertainment. This industry has become one of the largest entertainment industries globally, with an increase from $101 billion in 2016 to an expected $129 billion in 2020 (McDonald, 2017). However, while both environments are attractive advertising surroundings, both have a potential common hazard: they are often context-incongruent to a broad range of brands. Thematic congruence is often regarded as one of the most important criteria for media selection and related advertising success (Dahlen et al., 2008; Shamdasani et al., 2001). The results of previous studies indicate that incongruent ads or sponsorship efforts may generate feelings of frustration and ultimately lead consumers to ignore the sponsor’s message or advertisement (Halkias and Kokkinaki, 2014; Meyers-Levy and Tybout, 1989). Research, thus, has well documented the positive effects of ad-context congruency regarding the role of congruency for advertising success in different media environments (e.g., traditional media, web or videogames). Thematic congruency generates positive attitudes towards the advertised brand (Choi and Rifon, 2002), positive effects on ad perception (Cho, 2003), positive reactions to the ad (Moore et al., 2005) and stronger purchase intentions (Jeong and King, 2010). For instance, Chang et al. (2010) propose that for in-game advertising, it is important that congruency and integration between the context of the game and the advertisement are given to generate positive outcomes. Summing up, congruity between ads and content seems to foster positive advertising outcomes (Bishop et al., 2015; Halkias and Kokkinaki, 2013), while incongruency might have a negative impact on attitudes towards the ad (Lee and Mason, 1999) and on product evaluations (Meyers-Levy and Tybout, 1989; Segev et al., 2014).

© Springer Fachmedien Wiesbaden GmbH, part of Springer Nature 2020 F. Nimmermann, Congruency, Expectations and Consumer Behavior in Digital Environments, Handel und Internationales Marketing Retailing and International Marketing, https://doi.org/10.1007/978-3-658-28421-3_3

30

3 Essays

Hence, the research question arises, what opportunities do brands have to enhance advertising perception and to support consumers in resolving incongruity? According to schema theory, information that consumers received and which do not match preexisting knowledge and thus, cannot be processed in relation to preexisting knowledge, lead to a perception of incongruency (Mandler, 1982). However, research shows that cognitive flexibility helps to find possible connections within the given knowledge (Jhang et al., 2012) and thus, can enhance the resolution of incongruence, for instance, in the case of brand extensions (Kim and John, 2008; Meyers-Levy and Tybout, 1989). Here, construal level theory (CLT) postulates that psychological distance influences consumers’ mental degree of abstraction (i.e., their construal level); thus, the greater the psychological distance (e.g., time distance between own reference point and distant event) , the higher the construal level is and their level of abstraction (Trope et al., 2007). Hence, we argue that brands in incongruent advertising environments could use the findings regarding the effects of psychological distances to enhance the perception of their advertisement. Surprisingly, the impact of psychological distance as an anchor (e.g., in time) for advertisements has not been analyzed yet. Transferring the basic assumptions of schema (incongruity) theory (Mandler, 1982) to the context of our study, we assume that incongruent advertisements, anchored to a high level of psychological distance, should be processed by consumers with a higher level of abstraction, thus, with more ease and a higher level of perceived congruency, which in turn should positively influence consumer behavior. With this study, we contribute in numerous ways to knowledge in the field of advertising and consumer behavior, especially in the context of advertising in streaming services and mobile gaming apps: Based on the assumptions and implications of the CLT we investigate the impact of psychological distance on advertising processing and consumer-related outcome variables. We systematically manipulate context congruency (congruent, moderately congruent and incongruent) and the psychological distance (distant vs. near time and space anchors) in three experimental studies. We focus our discussion on a specific type of advertising environment, and we regard streaming services and mobile games as very important examples for digital advertising ecosystems because of the significant investments of businesses in this part of the advertising industry (Foye, 2018). We thereby expand the findings of previous studies in the field of advertising research by showing (1) how the level of mental abstraction influences the perceived congruency of brands and their advertising within a given digital context, (2) how the level of mental abstraction influences the overall advertisement evaluation, and (3) by investigating the impact of psychological distance on these relationships. Hereby, we derive implications for formulating strategies for

3.1 Impact of Mental Construal on Advertising Processing

31

brand placement in digital environments via sponsorship or advertisement in incongruent settings. 3.1.2 Literature Review and Hypotheses Development Conceptual Model

Our conceptual model (see Figure 3.1-1) builds on congruence theory, schema theory and construal level theory. In general, we believe that the level of context congruity has an impact on brand and advertising perception and on consumer behavior outcomes such as interest in more information. However, we suppose that psychological distance moderates the overall evaluation of contextual advertising and brand congruence as well as the relationships between the level of congruence and the outcome variables under review. • Ad/Brand Context Congruency • Ad Credibility • Ad Implementation Psychological Distance (Construal Level) Level of thematic BrandCongruence (Incongruent, Moderate Congruent & Congruent)

• • •

Attitude toward the Brand Attitude toward the Ad Interest in more Information

Controls: • • •

Brand and Product Involvement Context Expertise and Knowledge Demographics



Figure 3.1-1. Conceptual Model

Effects of Congruency Thematic congruence between an advertised brand and the overall context is often interpreted as one of the most important criteria for media selection (e.g., King et al., 2004; Moorman et al., 2002). Congruence refers to a matching of or a similarity between objects, i.e., objects are consistent with each other or at least have a structural correspondence (e.g., Maille and Fleck, 2011; Voorveld and Valkenburg, 2014). In the related research, the concept of fit is often used as synonym for congruence (Speed and

32

3 Essays

Thompson, 2000). Following congruence (congruity) theory (Osgood and Tannenbaum, 1955), individuals value the harmony of the given combination in general or of specific elements of combined objects (e.g., context and ad). Thus, context congruence is important when analyzing media context (Maille and Fleck, 2011). For instance, congruence between banner-ads and the content of a website lead to more positive attitudes towards the ad, the website and actual consumer behavior, such as the intention to click on the ad or to purchase the advertised brands (e.g., Moore et al., 2005). Very similar results were observed with regard to thematic congruence between ads and magazine topics (e.g., Moorman et al., 2002). For example, a match between the ad and the context of the magazine, e.g., a car ad in a car magazine, leads to a higher level of advertising effectiveness and positively affects brand attitudes and associations. However, few research in this context also found opposing effects and thus, positive effects of incongruency on central outcome variables which might be explained by the level of incongruency (e.g., Dahlen et al., 2008). Nonetheless, the majority stresses that ads that are perceived as contextual have been shown to exert greater immediate effects on consumers’ responses to the advertised brand or product than non-contextual ads (Guitart and Hervet, 2017). Moreover, for the case of video embedded ads, incongruent ad context improves ad memorability (Furnham et al., 2002). Perceived congruence, however, exerts positive effects when ads are directly embedded in the actual content, such as mid-roll video advertisements with a more concrete connection to the content (Li et al., 2014). For in-game advertising, Chang et al. (2010) show that it is important that the genre, which shapes the context of the game, matches the included advertisements in order to produce a positive behavioral response. Thus, while memory effects may be enhanced by placing brands in incongruous game settings, this may negatively influence brand attitudes (Lee and Faber, 2007). Generally, the sponsorship literature shows that higher levels of perceived congruence between a sponsor and the sponsored event evoke a positive image transfer, i.e., transferring positive attributes from the event to the brand perception (e.g., Mazodier and Merunka, 2011; Zdravkovic and Till, 2015). These results of previous research imply that if brands do not have natural links to the sponsored objects, marketers have to strengthen these links, e.g., by explaining the relationships or even by creating them artificially (Cornwell et al., 2006). Contextual advertising thus seems to be more effective than incongruent advertising. In contrast, a high level of incongruence and the need for strong cognitive adjustments lead to frustration of consumers and negative evaluations (d Astous and Bitz, 1995). In sum, congruence might be one of the most powerful predictors of persuasion (Mazodier and Quester, 2014) and even so positive effects of incongruency on e.g., memory have been found (e.g., Dahlen et al., 2008; Moorman et al., 2002), traditional views of persuasive congruent communication should not be entirely replaced by incongruent advertising (Halkias and Kokkinaki, 2014).

3.1 Impact of Mental Construal on Advertising Processing

33

The effects discussed above are likely to be impacted by a number of consumer characteristics, such as the relevance of ads and products, i.e., involvement, expertise and general relevance (e.g., Bishop et al., 2015; Lee and Faber, 2007; Verberckmoes et al., 2016). For instance, highly involved consumers may be more motivated to process incongruence because they are highly engaged in elaborate product evaluation (Petty and Cacioppo, 1986). In this context, De Pelsmacker, Geuens, and Anckaert (2002) show that context–advertising incongruence (e.g., in television or magazines) has a positive influence on advertising effectiveness for respondents with high involvement. Hence, highly involved consumers should be more likely to resolve brand–advertising incongruence (Mazodier and Merunka, 2011). In addition to these aspects mainly rooted in congruity theory, schemata congruence literature provides further explanations with regard to variables that might mediate or moderate the relationship between congruence or incongruence and related advertising effects (Mazodier and Quester, 2014). Researchers have observed strong relationships between congruence and schemata (e.g., Halkias and Kokkinaki, 2014; Torn, 2012; Verberckmoes et al., 2016). Literature stresses the importance of schemata because of their impact on the evaluation of congruence due to consumer expectations (Mandler, 1982; Stayman et al., 1992). Following schema theory, it is schemata that allow consumers to efficiently encode, store, and decode the information they encounter and the processing of incoming information follows certain patterns (Anderson, 1976). When consumers face new experiences, present schemas are triggered and new information is stored in relation to the existing patterns (Beals, 1998). Mandler (1982) argues that if the received information fits into consumers’ existing knowledge structures, this information is ‘schema congruent’ (e.g., a jeans manufacturer ad within a fashion website). However, if the information cannot be completely resolved by available patterns, the information is ‘schema-incongruent’ (e.g., the same jeans ad within a racing game). Thus, during the resolving process, which occurs when information does not fit into prior existing schemas, consumers must use greater effort to decode incoming information (Houston et al., 1987; Jeong and King, 2010). In contrast, when information is congruent with expectations, consumers can engage in simpler and easier processing (Sujan et al., 1986). Thus, research has exposed additional consumer characteristics that play a role in the overall advertising processing regarding congruence, but it has surprisingly neglected factors that influence the role of overall schemata that help consumers resolve negative effects of incongruence. For instance, studies in the context of product evaluation show that increasing analogical reasoning relates positively to product and brand congruence (e.g., emphasise benefits of the unusual composition) (Jhang et al., 2012). Furthermore, by triggering congruence, such

34

3 Essays

as by articulation (e.g., “proud sponsor since” or general explanations), consumers’ evaluation of the alignability of two objects (e.g., event and sponsor) can be enhanced (e.g., Olson and Thjømøe, 2011). Thus, if ads are atypical in regard to existing schemata and thus mismatch consumers’ expectations (e.g., Stayman et al., 1992), they should be met by broader categories within consumers’ schemata to support finding a match at all. Therefore, mental abstraction, which can be explained using psychological distances based on construal level theory, would be beneficial. Construal Level In a low-level construal, consumers process information on a more concrete level and are focusing their environment in great detail (e.g., a tree with all its branches and leaves). In a high construal, consumers are looking at the “bigger picture” rather than focusing on details (e.g., seeing the whole forest and not each tree for itself). The construal level is directly connected to psychological distance, with high distance evoking high levels of mental abstraction and vice versa (Trope and Liberman, 2010). Psychological distance is egocentric: “Its reference point is the self, here and now, and the different ways in which an object might be removed from that point — in time, space, social distance, and hypothetically” (Trope and Liberman, 2010, p. 440). Accordingly, with greater temporal or geographical distance, a stimulus cannot be directly experienced and is thus psychologically distant (Trope et al., 2007). Thus, greater temporal or geographical distance results in a higher construal level for the consumer, while a lower temporal or geographical distance leads to a lower construal level (Trope and Liberman, 2010). However, each dimension of psychological distance leads to an equal increase or decrease in total psychological distance (e.g., time distances do not mandatory lead to a higher psychological distance than hypothetical distances) (Trope and Liberman, 2012). In this context, Kim and John (2008) have shown that mental construal might play an important role when consumers form their perceptions of fit-importance between the brand and the extensions. However, here mental construal has only been studied as an a priori factor influencing consumers, while any impact on schemata regarding the processing of a perceived congruence has been neglected. Hypotheses In general, we propose that, following construal level theory, when advertisers use different anchors (temporal and/or spatial) before placing an advertisement, different levels of psychological distance are likely to be triggered. For example, if advertisers use an announcement that refers to time-related and/or geographic distance before they place an ad (e.g., introducing a distant upcoming event), this should evoke a high construal level, while a near anchoring should lead to a low construal level. In addition, the positive effects of ad-context and thematic congruency and their positive impact on consumer attitudes and behavior towards the ad and advertised brands have to be taken

3.1 Impact of Mental Construal on Advertising Processing

35

into account in the context of our study. Based on previous reasoning, we conclude that incongruent information creates a challenging task for consumers, representing a puzzle that consumers need to solve. A more thorough processing leads to enhanced confidence in “resolution judgments”, which is understood by consumers as a greater liking of the ad and the brand and advertising credibility (Lee 2000, p. 55). On the basis of our literature review, we conclude that thematically congruent brands (e.g., a watch manufacturer ad in a watch magazine) provide attributes that are directly related to preexisting category knowledge. Consumers are able to immediately understand an ad via existing associations with regard to the related context in which the ad is included. On the other hand, incongruity requires forming new associations within these categories (Jhang et al., 2012). While these aspects are more or less common knowledge, in our analyses, we focus on the relevance of construal level for these relationships. The construal level affects the categorizing abilities of consumers. A high construal level evokes broader but fewer categories, whereas a lower mental abstraction level leads to narrower but more numerous categories (Liberman et al., 2002). More precisely, consumers in a high construal level create more abstract and noticeably broader groups to categorize a given input. Transferring this general idea to schemata, we argue that a high construal level supports consumers in finding easier and even more numerous connections within these categories because they consider the “bigger picture” (Trope and Liberman, 2012). Consequently, a more distant anchor might lead one to categorize the advertising information into fewer, broader categories. Furthermore, research has shown that when consumers include a temporal orientation, i.e., future thinking, this is likely to impact their creativity in imagination (Chiu, 2012). Hence, in the case of a higher level of psychological distance, the context becomes more abstract, and it is therefore easier for consumers to categorize the thematically incongruent advertised brand and/or the ad into more diffuse contexts. Consumers in higher construal levels make decisions on the basis of abstract information, while consumers in lower construal levels promote concrete information (Braga et al., 2015). Moreover, consumers connect non-alignable information to a more abstract mindset (Malkoc et al., 2010). Therefore, higher-order information is preferred in the case of high-level construals, whereas concrete attributes are more effectively processed in the case of low-level construals (Hernandez et al., 2015). Following this reasoning, we suppose that more-concrete information should be more positively recognized in the context of thematically congruent brands and with a concrete level of mental abstraction due to the direct effects of congruence on the overall perception of brand and advertisement, whereas a more abstract construal level supports the resolution

36

3 Essays

of incongruence helps the individual recognize more positively abstract information. Finally, based on previous reasoning, we argue that a more successful processing of ad incongruity has positive effects on consumer perceptions, attitudes and behavior. We therefore expect a connection between the context in which advertisements are placed and ad effectiveness. However, we suppose that different construal levels impact the role of consumers’ congruence perception, i.e., congruity between the ad and the advertising context, in building positive consumer attitudes towards the brand and its advertising and in forming consumer behavior. Based on this theoretical reasoning, we strive in our analyses to answer the following two general hypotheses: H1: The psychological distances moderates the perceived contextual brand and advertising congruence such that (a) a thematically incongruent brand and advertising is enhanced in case of high psychological distance (i.e., high construal level) and (b) a thematically congruent brand and advertising perception is enhanced in case of low psychological distance (i.e., low construal level). H2: The psychological distance positively moderates the effect of congruency on the attitude toward the brand and consumer behavior such that (a) the effect is enhanced by a high psychological distance in the case of a thematically incongruent brand and (b) the effect is enhanced by a low psychological distance in the case of a thematically congruent brand. 3.1.3 Experiments To verify the proposed hypotheses, we conducted two consecutive experiments that built on one another: The intent of the first study was to gain first insights into whether the impact of an individual’s construal level on advertising perception holds true in a ‘mature’, but digital advertising environment. We therefore used a soccer stream as a contextual background for study 1. The second study was set in an eSports context and was conducted to gain insights into whether the degree of mental abstraction generated by an artificially implemented psychological distance impacts congruence perceptions and helps resolve incongruence, which in turn should affect consumer behavior relevant outcomes. Study 3, set in the context of mobile racing games, gives further insights into the effects of mental construal on the perception and evaluation of the advertisement. The intent behind the changes in setting (soccer stream vs. eSports streams vs. mobile games) was to analyze whether the results remain stable in different advertising environments, especially in digital media. 3.1.3.1 Study 1 Method and procedure

3.1 Impact of Mental Construal on Advertising Processing

37

This first study was set in the context of German soccer, a mature advertising environment where a high level of incongruence between advertised brands and the advertising environment is the status quo. Study 1 was intended to test the effects in this established environment, because we suppose that if the effects that are observed in an environment in which such incongruence occurs on regular basis, the hypothesized relations should be supported generally. Hence, to test whether mental construal impacts consumers’ fit perception, we conducted an online experiment by means of a 3x2 between-subjects design with the factors ‘thematic brand congruence’ (incongruent vs. moderately congruent vs. congruent) and ‘construal level’ (low vs. high). We used a manipulated video of a real soccer broadcast of the German Football League in all conditions. The subjects had to put themselves into the situation of watching this scene on television. The video started by showing players in a regular match. After nine seconds, one team passed the ball back to their goal keeper. During this pass, the video changed to a picture-in-picture view, showing the actual footage of the game being played in just in the upper right corner (approximately 20% of the overall frame size) and a neutral picture of a soccer field as the background. The rest of the screen was used to present the stimuli. The first factor, the brand congruence level, was manipulated by showing a different type of brand in each condition (incongruent vs. moderately congruent vs. congruent) with its specific ad, accompanied by the message that the brand supports the current game. We included three brands in our study to comply with the research recommendation to respect a congruity–incongruity continuum and to go beyond a dichotomous operationalization of incongruity, i.e., congruent versus incongruent (Halkias and Kokkinaki, 2014; Kim and John, 2008; Torn, 2012). Based on a pretest (N = 46, Mage=25.67, SD=8.04, 43.5% women) among participants similar to those in the main study, we chose Nike as a congruent brand because of its strong connection to soccer and sports in general. Coca-Cola was chosen as a moderately congruent brand because of its ubiquity in sports advertising and its role as a well-known brand in a broad range of advertising contexts. As an incongruent brand, we included Bosch (i.e., a power tool and household appliance manufacturer), which has no clear direct connection to soccer. The design, i.e., background and colors of the ads, were similar for all treatments. Only the company logo, the advertised product, and the claim in the ad were varied. During the advertisement, the kickoff of the keeper went straight offside to ensure that no “exciting” scene distracted the study participants from the actual advertising. After that, the video changed back to a full-size view of the actual game and faded out after five additional seconds. We opted for this realistic overlaybased advertisement style because of its wide usage in real sports broadcasts (e.g., Formula One). Moreover, using a real video helps the participants to put themselves into a real consumption situation of watching a sports broadcast. Following Slepian et al.

38

3 Essays

(2015), the second factor, construal level, was measured using a shortened form of the multi-item Behavior Identification Form (BIF) based on the work by Vallacher and Wegner (1987). Respondents were asked to report their construal levels using the BIF. Participants had to describe an activity (e.g., “locking a door”) by choosing an option that represented the action abstractly (“securing the house”) or concretely (“putting a key in a lock”). A low sum represented a concrete level of abstraction, while high values represented a more abstract construal level. This scale has been shown to be a valid and reliable measure of construal levels (Hernandez et al., 2015). Following Kim and Roedder John (2008), to identify two levels of mental construal, a median split was used. Hence, individuals scoring 4 or above (sum of all answers) were classified as “high” construal; individuals with scores of 3 or below were classified as “low” construal (7 items, 0 = concrete description – 1 = abstract description). Both groups were nearly equal in distribution (58% low construal levels). The participants in the study were recruited through social media websites, sportsrelated forums and university newsletters. N = 228 undergraduate respondents and random family members and friends participated in study 1 (n > 30 in each condition; Mage = 31.83 SD = 12.83, 136 women). The participants were randomly assigned to one of the three congruence conditions of the online experiment. After they read the introductory page, the respondents initially answered a question on their brand knowledge to ensure an existing fit perception regarding the brands used in the soccer broadcast (“Which of the following brand(s) are you generally familiar with?”). Before the actual experiment, the attitude towards advertising in general was measured by applying a seven-point scale based on Pollay and Mittal (1993) (3 items, e.g., “Generally, I like advertising." 1 = I totally disagree – 7 = I totally agree, Cronbach’s α = .874). Thereafter, the respondents received the following introduction to the experiment: “On the following page, you will see a segment of a soccer game. We kindly ask you to watch the video very attentively until the end. Before you start the video, please imagine that you have followed the match from the start and that the following scene now appears.” The “next page” button appeared after the video had finished. The participants were able to watch the video as long and as often as they liked, i.e., without any time restrictions. This procedure was chosen to minimize a distortion of the results due to a systematic bias in the experimental design (Bailar et al., 1977). After the video prime, we measured the individuals’ construal level using the shortened BIF scale (Vallacher and Wegner, 1987). Moreover, the perceived fit of the brand advertisement within the present context was measured using semantic differentials, as introduced by Becker-Olsen (2003), on a seven-point scale (6 items, e.g., 1 = does not make sense – 7 = makes sense, α = .93). Moreover, age, sex and education level were collected as controls. For a detailed overview in regard to the measures used in this essay, see appendix Table 3.1-7

3.1 Impact of Mental Construal on Advertising Processing

39

Results and Discussion First, the results show that our manipulation of congruence between ad and advertising context was successful (F(2,222) = 165.80, p < .001). Nike shows the highest level of congruence, while Bosch is perceived as an incongruent advertising brand in the case of soccer games. Moreover, Coca-Cola obtains the intended medium level of congruence evaluation. We were unable to observe a significant main effect of the construal level (F(1,222) = .009, p = .924), which stresses the importance of the finding “right” match between construal and congruence level. As expected, we can show that while the expected manipulation is generally confirmed, the perceived fit varies in the three congruence conditions with differences in the individual’s construal level (F(2,222) = 4.132, p = .017). Table 3.1-1 depicts these results. Regarding the attitude towards advertising in general, we do not observe any effect (F(2,222) = 1.315, p = .271), which supports that this attitude does not affect the present results. Moreover, there is no significant impact of our control variables age, sex, and education level on the variables under review in this study. Table 3.1-1. ANOVA Results (DV = Congruence)

Brand fit post-experiment Factors

Congruency (CG)

F

Construal Level (CL)

High

Moderate

Low

CG

CL

CG ✕ CL

Low construal level

6.31 (.83)

3.59 (1.47)

2.16 (1.06)

High construal level

5.90 (1.08)

3.26 (1.52)

2.84 (1.53)

165.80 ***

.009 n.s.

4.132 ** η²=.036

Post-perceived brand fit towards context, 1 = low perceived congruence, 7 = high perceived congruence (* .05 < p < 0.1; ** p < .05; *** p < .001; n. s. = not significant); Mean (SD in brackets)

To determine which state of mental construal produces significant differences within the three levels of congruency, we used multiple t-tests where we successively selected each level of congruency and compared the two construal levels. In case of the congruent brand, we find a significant difference regarding the difference in fit evaluation between a low and high construal level at a p < .10 level (tcongruent(80)=1.819, p = .074). However, a closer look at the mean values reveals a significant difference in the case of low congruence (tincongruent(67)=2.188, p = .042). We do not observe any significant difference in case of moderate incongruence (tmedium_congruent(77)=.939, p = .351). The results generally support our assumption, that mental construal impacts the perception of congruency between the ad and the context. Here, due to a match between

40

3 Essays

incoming information and mental construal (Braga et al., 2015), it seems that in the case of a more concrete mindset, i.e., a low construal level, a brand with a high level of congruence is more positively processed, whereas more-abstract, incongruent information seems indeed to be preferred in the case of high construals. Consumers’ categorization flexibility, which is evoked via high construal (Liberman et al., 2002), seems to support the positive impact on fit evaluation in case of incongruence. Nonetheless, in the case of medium congruence, we do not observe any significant effect. However, it seems that medium congruence is more suitable in the case of a low construal level because of the more obvious links, in contrast to full incongruence (see Figure 3.1-2). Nonetheless, in study 1, the construal level is interpreted as a predisposition of consumers. Perceived Advertising Congruence 7

p=.074

6 5 4

p=.351 3

p=.042 2 1 High CG

Medium CG Low CL

Low CG

High CL

Figure 3.1-2. Inner-condition Differences of Advertising Congruence Perception, CG = congruence, CL = construal level

3.1.3.2 Study 2 Method and procedure To deepen insights on whether different psychological distance anchoring lead to different mental abstraction levels, in another advertising environment, we performed our second online experiment among German consumers within the context of streaming services, more specifically, in an eSports streaming environment. We used a 3x2 between-subjects design with the factors ‘thematic brand congruency’ (incongruent vs. moderately congruent vs. congruent) and ‘psychological distance’ (low vs. high). In all conditions, we used a manipulated video of a real eSports broadcast from ESL ONE Cologne (see Figure 3.1-3). Respondents were asked to put themselves into the situation of watching this scene in a real broadcast within a streaming service.

3.1 Impact of Mental Construal on Advertising Processing

Intro Intro

Prime Prime

41

Brand Brand

Outro Outro

Figure 3.1-3. Schematic flow of the video prime in study 2

The video started by showing players in close-up and then switched to in-game footage showing the athletes’ warm-up exercises. During this transition, the video changed to a picture-in-picture view, showing the actual footage of the game being played in the upper right corner (approximately 20% of the overall frame size), with the rest of the screen used to present the stimuli: information about the current or an upcoming event varying within the treatments in accordance with the research design. We manipulated the first factor, psychological distance, by means of temporal and spatial distance anchors. In the low psychological distance condition, the prime contained information about a “local event” (low spatial distance) and that the next match begins in “a few seconds” (low temporal distance). Based on the research of Hernandez, Wright, and Rodrigues (2015), the screen also contained concrete information (attributes) about the next match (map, price, number of players and overall maps). This information was used to evoke a more concrete mental construal of information processing and to give a clearer picture about the event. In the high psychological distance condition, the prime contained information about a yet unspecified event that would occur “in South America” (high spatial distance) and would start “within the given year” (high temporal distance). Afterwards, the participants were asked whether they would like to take part in the upcoming event, “when professional teams will compete against each other for a large prize pool”. Beyond that, no concrete information was given. Then, the information faded out, and a brand-related ad was shown for six seconds in the former information area to ensure that the respondents had enough time for a detailed ad perception. The second factor, the congruence level, was manipulated by showing a different type of brand in each condition (incongruent vs. moderately congruent vs. congruent) with its specific ad. Based on a pretest (N = 62, Mage=26.90, SD=7.09, 33 women) with participants similar to those in the main study, we chose Logitech as the congruent brand because of their eSports-related hardware. As in study 1, Coca-Cola was again chosen as the moderately congruent brand because of their advertisement ubiquity. Even though Coca-Cola has no direct link with eSports, it is a well-known brand in a broad range of advertising contexts. As the incongruent brand, we choose Levi’s, which has been completely absent from eSports advertising. The design, i.e., the

42

3 Essays

background and colors of the ads, were similar for all treatments. Only the company logo, the advertised product, and the claim in the ad were varied. After that, the video changed back to a full-sized view of the in-game footage and faded out after five additional seconds. Participants for this study were recruited via social media websites, eSports-related forums and university newsletters. N = 214 undergraduates participated in the first study (n > 32 in each condition, Mage = 25.59, SD = 4.97, 46,7% females). The participants were randomly assigned to one of the six experimental conditions. After the introductory page, the respondents initially indicated brand knowledge via one question. As a manipulation check, we evaluated their congruence perception regarding the brands used in the video and the eSports environment. Attitude toward the brand was measured via a scale based on Stuart et al. (1987) (5 items; e.g., 1 = not appealing – 7 = appealing, Cronbach’s α = .80). Perceived brand and advertising congruence with the present context (i.e., the eSports stream) were measured using semantic differentials as introduced by Becker-Olsen (2003) on a seven-point scale (6 items, e.g., 1 = does not make sense – 7 = makes sense, brand-congruence α = .95, advertisement-congruence α = .96). Moreover, respondents were asked whether, prior to this study, they had seen an eSports broadcast. In addition, respondents’ involvement and expertise regarding eSports was documented. Expertise in eSports was measured using a seven-point scale based on Griffith et al. (2004) (3 items; e.g., “I know very much about eSports” 1 = I totally disagree – 7 = I totally agree, α = .92). After they answered questions regarding the constructs mentioned above for all brands, the respondents were introduced to the experiment: “On the following page, you will see a segment of an eSports stream. We kindly ask you to watch the video very attentively until the end. Before you start the video, please imagine that you have followed the match from the start and that the following scene now appears.” The “next page” button appeared after the video had finished. The participants were able to watch the video as long and as often as they liked without any time restrictions. The intent was for the participants to have a satisfactory impression of the stream footage. Moreover, this procedure was chosen to minimize distortion of the results due to a systematic bias in the experimental design (Bailar et al., 1977). As a manipulation check for the construal level, the subjects were asked to report their construal levels, again via the shortened multi-item BIF. Afterwards, the respondents had to answer questions regarding the main constructs in order to measure the relative change in congruence perception and attitude. Lastly, the respondents had to recall the brand they had seen and what kind of information was given within the video (next match vs. match in one year). Age, sex and education level were documented as control variables.

3.1 Impact of Mental Construal on Advertising Processing

43

To test whether the manipulation of the construal level by the given information of the current event (low temporal and spatial distance) vs. the event next year (high temporal and spatial distance) was successful, we compared respondents’ answers on the BIF scale in the two corresponding conditions. The results showed that the manipulation of the construal level was successful: Respondents who saw the information about the current event reported a BIF score of Mlow = 2.95, SD = 2.03. Respondents who were shown the distant event reported a significant higher construal level (Mhigh = 3.58, SD = 2.11, F(1,212) = 4.872, p = .028). Following Baron and Kenny (1986), the BIF score (effect of the psychological distance) has to be independent from the congruence level of the brand, which our results confirm (F(2,211) = .104, p = .902). Moreover, the results show that the perception of brand and advertising congruence depends on the brand category and the related products (see Table 3.1-2). As intended via our manipulation, Logitech shows the highest level of congruence, while Levi’s is interpreted as thematically incongruent pre-experimental manipulation. Coca-Cola is perceived as moderately congruent. Based on these results, we argue that the manipulation of the congruence levels was successful. Moreover, as expected, the pre-experiment congruence does not depend on the construal-level manipulation and therefore shows no interaction effect with the manipulated congruence level (Fbrand_congruence(2, 208) = .348, p = .706 & Fad_congruence(2, 208) = .270, p = .764). Table 3.1-2. MANOVA Results – Pre-Experiment Brand and Advertising congruence Factor: Congruence

Brand congruence F(2,211)=87.170***

Advertising congruence F(2,211)=54.056***

Logitech (high congruence)

4.88 (1.35)

4.49 (1.18)

Coca-Cola (moderate congruence)

3.91 (1.50)

3.50 (1.72)

Levi’s (low congruence)

2.08 (1.12)

2.19 (1.14)

Pre-perceived brand and advertising congruence with context, 1 = low perceived congruence, 7 = high perceived congruence (* .05 < p < 0.1; ** p < .05; *** p < .001; n. s. = not significant); Mean (SD in brackets)

Results and discussion To test our hypotheses, we performed a MANOVA with regard to perceived levels of congruence between the brand and advertising and the eSports environment after the video stimuli was shown. MANOVA results show that the construal level and the level of congruency interact and that with regard to both aspects of brand and advertising congruence, the psychological distance, in which the consumer evaluates the context congruence, matters (see Table 3.1-3).

44

3 Essays

Table 3.1-3. MANOVA Results (DV = Congruence)

Post-experiment Brand Congruence Factors

Congruency (CG)

F

Psychological Distance (PD)

High

Moderate

Low

Low psychological distance

5.56 (1.36)

3.50 (1.63)

2.01 (1.18)

High psychological distance

5.08 (1.50)

4.11 (1.71)

2.74 (1.85)

CG

PD

CG ✕ PD

66.391 ***

1.800 n.s.

3.185 ** η²=.031

Post-experiment Advertising Congruence Factors

Congruency (CG)

F

Psychological Distance (PD)

High

Moderate

Low

Low psychological distance

5.20 (1.34)

3.30 (1.77)

2.03 (1.24)

High psychological distance

4.45 (1.58)

3.97 (1.88)

2.89 (1.89)

CG

PD

CG ✕ PD

38.413 ***

1.294 n.s.

5.077 ** η²=.047

Post-perceived brand and advertising congruence with context, 1 = low perceived congruence, 7 = high perceived congruence (* .05 < p < 0.1; ** p < .05; *** p < .001; n. s. = not significant); Mean (SD in brackets)

To determine which state of mental abstraction within the three general levels of congruency produces significant differences, we used multiple planned contrast tests where we successively selected each level of congruency and compared the two mental abstraction levels within each congruence level. In the case of brand congruence, the results do not show any differences in the high level of congruence condition (t(64)=1.354, p=.180). In this case, the abstraction level at which the consumer of an eSports stream perceives the brand does not seem to matter. A nearly identical result can be observed in the case of medium congruency (t(64)=1.491, p=.141). In contrast, in the condition with a low level of congruence (t(80)=2.132, p=.036), a high level of psychological distance causes consumers to perceive a higher level of context congruence. For the actual advertising, our results confirm a significant difference for the distinct psychological distance levels for two out of the three congruence conditions and, therefore, an impact of the mental abstraction levels. In the case of high congruence (t(64)=2.088, p=.041), the more distant psychological state even negatively impacts the post-treatment congruence, in contrast to the pre-advertising evaluation. In the cases of

3.1 Impact of Mental Construal on Advertising Processing

45

moderate congruence (t(64)=1.475, p=.145) and low congruence (t(80)=2.454, p=.016), a high level of construal – as expected – produces a positive effect. Nevertheless, in the case of medium congruence, the mean difference is not significant. Because we collected congruence twice, i.e., pre- and post-experiment, we calculated the difference (Δcongruence) between brand-congruencepost_video (advertisingcongruencepost_video) and brand-congruencepre_video (advertising-congruencepre_video) and again applied planned contrast tests to determine whether the growth or decrease of congruence from pre- to post-measurement within a brand differs significantly due to the psychological distance (see Figure 3.1-4). ΔBRAND-FIT

decrease

growth

Low PD

ΔADVERTISING-FIT

High PD

Low PD

1

1

0

0

-1

High PD

-1 HIGH CG

MODERATE CG

LOW CG

HIGH CG

MODERATE CG

LOW CG

Figure 3.1-4. Delta-Congruence-Perception, CG = congruence, PD = psychological distance

In the case of Δbrand-congruence, we only find a significant difference in the low congruence level condition: thigh_congruence(64)=1.386, p=.171, tmedium_congruence(64)=1.543, p=.128 and tlow_congruence(80)=2.735, p=.008. However, the results show a significant impact of psychological distance in the case of Δadvertising-congruence in all three congruence levels: thigh_congruence(64)=2.912, p=.005, tmedium_congruence(64)=2.004, p=.049 and tlow_congruence(80)=2.931, p=.005. As expected, high-distance temporal anchoring in the case of low congruence leads to a significant growth of perceived congruence, while low-distance temporal anchoring is more favorable in the case of high congruence. In addition, we conducted two moderation analyses by using PROCESS for SPSS (Hayes, 2017) to reveal any potential moderation effects of the involvement and expertise regarding eSports that could influence the main effect of psychological distance. However, neither involvement in eSports nor eSports expertise influenced the perception of the advertising or brand congruence in the case of eSports streaming services in our data.

46

3 Essays

The results support H1a and b: Anchoring advertisements to different psychological distances leads to a different processing of perceived congruence. A high consumer construal level seems to support connections within the patterns and makes it simpler to find a link between the context and the advertising brand. This result is in line with previous research where future orientated temporal time frames lead to a more positive evaluation in case of incongruent brand extensions (Jhang et al., 2012). However, supporting a low level of mental construal due to a low psychological distance anchor might moderate perceived congruence negatively in the case of incongruent brands or at least show a more neutral effect. As consumers seek for a match between their construal level and the heuristics used (Braga et al., 2015), the mismatch might lead to less positive evaluations. In the case of a congruent brand, the results show the opposite effect, although the negative effect of a “wrong” construal level, i.e., a high distance anchor, is not crucial. In this second case, it seems more promising to advertise in a more concrete way, e.g., by referring to the psychologically closer anchor. Hence, in both cases the perceived congruence is positively moderated by the psychological distance. More precisely, the results show that only by referring the advertisement to a more distant (vs. close) setting in the same context, the distance moderates the context congruence such that the evaluation of congruency is positively strengthened. The attitude toward the brand does not differ significantly between high or low construal level conditions in this study. We suppose that the attitude toward the brand is a more consistent construct, and thus, an immediate change via construal level manipulation might be hard to achieve. In contrast, congruence is a present-based perception and is therefore more flexible. This might also explain why we only observe significant interaction effects for brand congruence perception in the case of incongruent brands. The overall brand perception might be more stable, whereas the perception of the congruence between the advertising and the context is more present-based and is therefore susceptible to distance anchoring. Nevertheless, with a higher congruence between ad and context, brand credibility might increase, and the brand will become more favorable (Dahlen and Lange, 2005). 3.1.3.3 Study 3 Method and procedure To gain a more profound insight into the effects of mental construal on advertising perception, we conducted a third online experiment among German consumers. We tested our hypotheses in this study again by means of a 3x2 between subject design as in the previous experiments, with the factors ‘thematic brand congruency’ (incongruent vs. moderately congruent vs. congruent) and ‘psychological distance’ (low vs. high). In this case, we conducted the study in a mobile video game setting to analyze whether the

3.1 Impact of Mental Construal on Advertising Processing

47

results hold in this ad environment. We used a manipulated video of a videogame trailer that showed in-game footage through direct screen capturing. Participants were asked to put themselves in the situation of playing the game as the scene depicted in Figure 3.1-5 appeared. Intro

Prime

Brand

Outro

Figure 3.1-5. Schematic flow of the video-prime in experiment 3

The video started by showing the screen for selecting a car. Thereupon, the car was selected, and the game menu changed to a racetrack selection where the player hit the button “Race”. The game changed to the real full-sized load-screen for three seconds, followed by a manipulated transition to a full-size information area for a related game add-on (i.e., downloadable content for the game), which we used as prime for 22 seconds. We again manipulated psychological distance by means of the temporal and spatial distance. In the low psychological distance condition, the prime contained information about a game setting called “One Night in Cologne”. Moreover, we added the information that the game is available today for “only 0.99 €” at a specific app-store. The window also contained concrete information (attributes) about game features (three new racing tracks in Cologne, two new car-paintworks, a new game mode and a new car). In the high psychological distance condition, the prime anchor contained information about a game add-on that would be released “still this year”, with an unspecified price, and would be available “everywhere”. The setting was displayed with the title “A Journey to Tokyo”. In contrast to the low psychological distance setting, the information area did not contain specific game attributes but rather the benefits of playing the game (e.g., new game modes for a joyful experience). Then, the information faded out and a brand-related ad was shown for six seconds in the former information area. The second factor, brand congruency, was manipulated as in the previous experiments. Based on a pretest (N = 58, Mage=23.66, SD=6.83, 55.2% women), we used Volkswagen to represent a congruent brand due to their strong proximity to cars. For comparability with study 1, we used Coca-Cola for moderate incongruence and Levi’s as an incongruent brand. The design, i.e., the background and colors of the ads, were equal in all conditions and similar to previous experiments. Again, only the brand logo, the advertised product, and the claim in the ad were exchanged. Moreover, we added a loading animation in the lower right frame for a more realistic impression that

48

3 Essays

the game was still in a loading segment. After that, the video changed back to a fullsized view of the in-game footage (i.e., the real loading screen for three additional seconds, followed by the racing start sequence). Afterwards, the video faded out. We chose this overlay-based style because there is a trend in video gaming practice to show additional information about the video game within loading screens or even offer mini games to tackle the waiting period (Campbell, 2015). The procedure in study 3 was analogous to the procedure in study 2, with three central differences: (1) we changed the setting to game-based advertising, (2) we measured the relevant outcome variables only once after the experiment, and (3) we focused in the third experiment on the effects on advertisement perception. The one-time measurement was based on the decision to analyze the moderating effects of psychological distance on ad and congruence perception without a possible bias of self-evaluation adjustment. N = 221 undergraduates participated in the third study (n > 31 in each condition, Mage = 26.46, SD = 16.91, 139 women). The structure of the questionnaire was identical: After the start page, the respondents initially answered a question regarding brand knowledge and were randomly assigned to one of the six experimental conditions of the online experiment. Then, the respondents completed measures related to several control variables, including involvement with mobile games, interest in mobile games and brand involvement. Involvement with mobile games was measured by applying a scale based on Segev et al. (2014) (4 items; e.g., “I think of myself as a mobile gamer” 1 = I totally disagree – 7 = I totally agree α = .85). Interest in mobile games was measured on a seven-point scale by Olney et al. (1991) (4 items; e.g., “Mobile games are interesting to me” 1 = I totally disagree – 7 = I totally agree, α = .95). Moreover, brand involvement was measured based on a scale introduced by Voss et al. (2003) (3 items; e.g., “I attach great importance to this brand” 1 = I totally disagree – 7 = I totally agree, α = .86). Afterwards, the respondents were introduced to the experiment, which was identical to study 1, but adapted for racing games. As a manipulation check for construal level, we again used the shortened multi-item BIF, complemented by questions regarding the perceived temporal and spatial distance. Afterwards, respondents had to answer questions regarding the main constructs, i.e., the perceived congruence of the advertisement within the present context, the perceived implementation of the ad, and the attitude toward the brand. The measurement of perceived advertising congruence was operationalized by using the same Becker-Olsen (2003) scale as in study 1 (α = .96). In addition, the perceived ad implementation was measured by adapting a scale based on Rifon et al. (2004) (3 items; e.g., “The advertisement is well integrated into the game” 1 = I totally disagree – 7 = I totally agree, α = .85). The attitude toward the brand was operationalized in a way similar to study 2, based on the scale by Stuart et al. (1987) (α = .94). Thereafter, the respondents had to answer questions regarding their attitude toward the ad, the ad’s credibility, brand involvement and product involvement.

3.1 Impact of Mental Construal on Advertising Processing

49

The attitude toward the ad was operationalized on a on a seven-point scale by De Pelsmacker et al. (2002) (6 items; e.g., “I found it attractive” 1 = I totally disagree – 7 = I totally agree, α = .81). The credibility of the advertisement was adapted based on a scale introduced by Pollay and Mittal (1993) (3 items; e.g., “In general, I feel that I can trust the shown advertisement” 1 = I totally disagree – 7 = I totally agree, α = .79). Finally, the respondents had to answer questions regarding their product involvement and mobile game expertise. Moreover, we measured the interest in seeking more information after the ad was shown to the respondents. The involvement with the product category was operationalized by applying a scale based on De Pelsmacker et al. (2002) (3 items; e.g., “[The product category shown in the video] … is important to me” 1 = I totally disagree – 7 = I totally agree, α = .86). Expertise in mobile games was documented based on a scale by Griffith et al. (2004), as used in study 2 (α = .81). Interest in seeking more information about the brand was adapted based on a scale introduced by Griffin et al. (2008) (6 items; e.g., “I'm likely to go out of my way to get more information” 1 = I totally disagree – 7 = I totally agree, α = .73). Finally, the respondents had to recall the brand they had seen and what kind of information was given within the video (next match vs. match in one year), followed by the last page where we gathered age, sex and education level as additional control variables. To test the success of manipulating the psychological distance by providing information about the released game add-on set in Cologne (low temporal and spatial distance) vs. an upcoming add-on set in Japan (high temporal and spatial distance), we compared the respondents’ answers regarding the perceived temporal distance (“The release of the expansion of Real Racing (i.e., the time when the expansion can be played) was...” 1 = very close – 7 = very far away) and the spatial distance (“The geographical distance of the expansion of Real Racing (i.e., the place where the expansion is set) was...” 1 = very close – 7 = very far away). Respondents with a low distance treatment reported mean values of Mlow_temporal=3.21, SD = 1.25 and Mlow_spatial=3.18, SD=1.41, while respondents in the high psychological distance treatment reported mean values of Mhigh_temporal=4.00, SD= 1.03 and Mhigh_spatial=4.58, SD=1.56. In both conditions, the distance perceptions differed significantly (Ftemporal(1,219)=26.292, p 24 in each treatment. More than 60% of our sample are shopping more than once a month online, which ensures knowledge about how to shop online and reviews in regard to products. In addition, across all treatments at least 50%

3.4 Online Complaints in the Eye of the Beholder

119

of the respondents purchased a product online within 30 days of conducting the survey. It can therefore be assumed that this purchase is still mentally present so that respondents could imagine themselves more easily into a purchase situation. 3.4.4 Results and Implications First of all, it can be shown that it matters how a company responds to a public complaint. That is, the choice of the response type by the company influences both the behavior and attitudes towards the company of the observers of public complaint handling. Table 3.4-2 shows the results of MANOVA conducted for the dimensions of an observer’s reaction (dependent variables) to different types of organizational response (independent variable). These results are significant with F = 1.534 and Wilk’sλ = .691 for p < .05. Table 3.4-2. Main effects of organizational response on observer’s behavior MS

F-Value

η2

Attitude

8.295

5.182***

0.152

Trust

7.285

5.052***

0.149

Liking

7.316

4.268***

0.129

Risk attitude

3.384

1.926*

0.063

Purchase intention

8.947

3.869**

0.118

Positive WOM intention

8.816

3.380**

0.105

Negative WOM intention

3.551

1.181 n.s.

0.039

Dimensions of observer’s behavior

(* .05 < p < 0.1; ** p < .01; *** p < .001; n. s. = not significant)

We can see that prospective customers are rather careful in making definite negative judgments based on a single observation of public complaint handling. Their assessment of risk and their intention to spread negative WOM do not differ substantially between various response types. This indicates that customers tend to give a company a “second chance” before making a decision to actively spread the “bad news”. Nevertheless, the same observation of a public complaint handling encounter can make them cautious, since it strongly impacts prospective customer’s attitudes towards the company, including their purchase intentions, trust in, and sympathy to the company. These results hint to a contemplation that although the attitudes and the intentions are strongly impacted by the company’s choice of complaint response, the damage of the first

120

3 Essays

encounter with that response remains rather limited to the recipient and will not be (at least for the first time) multiplied by them. A closer look to Table 3.4-3 that presents mean values of correspondent variables across different types of organizational response confirms this contemplation. Furthermore, they allow us to build a rank-order of different reaction types with respect to their potential to mitigate and even recover the negative effects of a public complaint.

Request of a direct contact

No response

Accuse customer

Apology

Accuse thirdparty

Deny fault

Redress

Apology + redress

Dimensions of observer’s behavior

Positive reviews

Table 3.4-3. Mean values of behavioral variables across different types of organizational response

Overall rank order

1

2

3

4

5

6

7

8

9

Attitude

5.68 (.81)

5.30 (1.04)

4.81 (1.23)

4.83 (.89)

4.39 (1.50)

4.67 (1.31)

4.53 (1.48)

3.99 (.99)

3.81 (1.38)

Trust

5.45 (.81)

4.53 (1.36)

4.59 (1.09)

4.32 (.74)

4.10 (1.33)

4.21 (1.31)

4.32 (1.59)

-

3.73 (1.23)

Liking

5.63 (.62)

4.97 (1.32)

4.42 (1.22)

4.52 (1.01)

4.36 (1.59)

4.43 (1.63)

4.43 (1.51)

4.18 (1.08)

3.84 (1.21)

Risk attitude

5.22 (.94)

4.50 (1.36)

4.42 (1.22)

4.59 (1.18)

4.67 (1.49)

4.44 (1.49)

4.19 (1.48)

3.90 (.96)

4.14 (1.25)

Purchase intention

5.71 (1.04)

4.85 (1.31)

4.78 (1.29)

4.53 (1.51)

4.58 (1.89)

4.56 (1.47)

4.40 (1.49)

4.07 (1.09)

3.72 (1.66)

Positive WOM intention

4.53 (1.17)

3.75 (1.71)

4.11 (1.29)

3.31 (1.63)

3.44 (1.86)

3.09 (1.73)

3.22 (1.48)

3.15 (1.39)

2.88 (1.69)

Note: Presented in the descending order of the potential to mitigate negative effects; values in parentheses are standard deviations; columns with a grey background represent the upper and lower benchmarks; only significant effects are considered.

First of all, it can be seen that the combination of an apology and redress evokes the most positive response in the behavior and attitudes of the observer of public complain handling. Remarkably, the effects are stronger than those in the case when the observer didn’t observe any negative information about the company. This fact is interesting from two perspectives. First, it is very similar to the service recovery paradox that is known from complaint management. Second, in the field of complaint management this combination of organizational response dimensions is also known to bear the most

3.4 Online Complaints in the Eye of the Beholder

121

positive behavioral response of a complainer (Davidow, 2003). Hence, the benefits of a company from this response type are twofold. This response both recovers a complainer’s negative experience and induces the most positive response in the eye of the beholder of a public complaint handling act. In the public domain it becomes even more important, since it is known that satisfied customers do not tend to spread positive WOM about the company. In this case, an observation of successful complaint handling takes the role of the positive WOM and compensates the latter effect. What’s interesting, however, is that being used separately from each other, the redress and apology response strategies lie relatively far apart from each other with respect to rank order (positions 3 and 6 correspondingly). T-tests reveal significant differences for attitude (t(46)=2.918, p=.006) and trust towards the company (t(46)=3.095, p=.003), as well as in case of purchase intention (t(46)=2.622, p=.012) and risk attitude (t(46)=4.305, p < .001). In other words, observers value an offer of a monetary compensation more than an apology, even though they do not personally benefit from it. A possible explanation for this circumstance can be that by reading a company’s responses to public complaints, observers try to evaluate their future risks in making business with the company, i.e., to asses the level of expected loss and/or compensation (both moral and monetary) in case of the company’s failure. Thus, observers feel themselves more secure with respect to future relationship with the company when the company demonstrates its readiness to repair the failure by offering monetary compensation. Based on the above, it seems reasonable for companies to offer monetary compensation when handling justified public complaints. However, monetary compensation always means costs for a company. Offering money in response to every complaint doesn’t seem to be a preferable complaint handling strategy. Moreover, it is possible that customers will uncover the reasoning underlying such a strategy and will try to misuse it to their own advantage. Nevertheless, from practical perspective, it can be argued that this strategy is at least worth consideration. Hence, we also encourage researchers to investigate into how often a company should offer monetary compensation in response to public complaints and/or what complaints deserve monetary compensation from an economic point of view. Other response strategies that companies can follow to counteract negative consequences of public complaints are accusing a third-party and denial of the fault (rank-orders 5 and 4). In our experiment, these strategies produce stronger overall recovery effects on observer’s behavior and attitudes than an apology. This observation might lead to a conclusion that when a company doesn’t want to provide a redress, the best strategy might be to deny the existence of the problem. At this point we have to

122

3 Essays

warn our reader not to make this mistake, since online communication isn’t restricted to a single request-response move. Companies should be aware that a customer can reply to their denial presenting her own point of view to the problem and the tides can turn against the denier caught in a lie. Hence, a denial of the problem and accusation of a third-party can only be efficient in recovering negative effects of public complaint handling when the company indeed is not responsible for the problem and can prove its point of view. Another very interesting observation is that the effect of a direct contact request as a response to a public complaint (e.g., to provide additional details on a problem) is even worse than not responding to a public complaint at all. Here the outcomes for a complainer and an observer seem to differ again essentially. The complainer will probably have her problem solved. The observer, however, at this point seems to be intrigued with the problem resolution. The absence of additional information about the resolution seems to cause internal cognitive tension in the observer. This unresolved tension seems to induce disaffection in the observer that, in turn, leads to even worse attitudes and behavioral intentions. Although these assertions are subject to tests during our upcoming study, at this point we argue that it makes sense for companies to provide a public feedback on a complainer’s problem resolution to ensure the redress in the eye of the beholder. Here, we appeal especially to telecommunication companies that showed a dominant tendency to employ “direct contact request” strategy in our preliminary study. Nonetheless, to get a more profound knowledge regarding the processing of the companies answer by the observer, we conducted multiple analyses of covariance (ANCOVA) for mediation analyses. The latter has been shown as reliable method for experimental designs to investigate potential mediations effects of given variables (Gorn et al., 2004; Hayes and Preacher, 2013; Pham and Muthukrishnan, 2002; Song and Zinkhan, 2008). Here, the experimental conditions are included as factor (i.e., independent variable), the behavioral and attitudinal outcomes as dependent variables and the potential mediators are gradually included as covariates. The decrease of mean square of the main effect after including the covariate indicates the mediation effect, as long as the mediator variable has a significant effect on the dependent variable (Baron and Kenny, 1986; Pham and Muthukrishnan, 2002). First, following Baron and Kennys’ causal steps (1986) we first investigated for direct effects of the experimental conditions on the potential mediators. As expected, results indicate a potential mediation of the satisfaction with the answer based on the expectations (F(6,173)=4.317, p=.001), the perceived fairness of the complaint answer (Finteractional(6,173)=4.223, p=.001; Fprocedural(6,173)=4.215, p=.001; Fdistributive(6,173)= 2.187, p=.046), the perceived credibility (F(6,173)=4.041, p=.001) as well as the

3.4 Online Complaints in the Eye of the Beholder

123

usefulness of the answer (F(6,173)=5.548, p. 05)

Results indicate strong mediating effects of the satisfaction with the answer based on expectations, the distributive fairness (e.g., monetary outcome) and especially, the overall credibility. In these cases, results show full mediation effects for several variables under review. However, the remaining variables at least show partial mediating effects. As predicted, the processing of the answer based on expectations (i.e., how the company should answer) and a corresponding satisfaction if these expectations are fulfilled is an important driver, not only for the complainant (de Matos et al., 2007), but also for the silent observer of the company’s answer. Moreover, the credibility of the company’s answer as information source is of particular importance, which also have been observed in case of the complainant (e.g., Breitsohl et al., 2010; Cheung et al., 2009; Sternthal et al., 1978). Hence, companies should be aware of offering meaningful, yet, realistic offerings and explanations regarding the actual complaint. Surprisingly, the perception of the distributive fairness is more important for the observer than the interactional fairness, which has been shown as more powerful in case

124

3 Essays

of the complaint (e.g., Blodgett et al., 1997). Still, it seems comprehensive that the complainant is more prone to the actual handling (e.g., a more sensitive treatment) because of vulnerability due to the service failure (Davidow, 2003), whereas consumers seeking for additional information prior to a purchase want to reduce their risk, so that the actual “outcome” in case of service failure seems more important (Hennig-Thurau and Walsh, 2004). Moreover, the actual cause attribution seems to be less important in comparison to the complainant (Laczniak et al., 2001). Here, one might argue that again regarding a risk reduction the actual cause does not matter, because the evaluation of the outcome is more important to get a glimpse of what would happen in an own case of service failure. The latter is moreover supported by the fact, that the actual risk attitude is not mediated by the cause attribution. Nevertheless, this places even greater emphasis on what the company responds to the complaint, regardless of the reason for the complaint. 3.4.5 Conclusion Summary of findings The results of our study show that the way a company responds to a public complaint is important. It seems reasonable for companies to follow an apology-and-redress response strategy. However, if this approach is not feasible for a company, it should apply an accuse strategy only when it can prove that it is not responsible for a complainer’s problem. Furthermore, the results indicate that a direct contact request should always be followed by a feedback on the problem’s resolution. Otherwise, companies risk losing more than they can gain by simply not answering a public complaint, or even accusing a customer. Nonetheless, results also offer first implications regarding the overall processing of complaint management by information seeking consumers. Here, results show that some of the already known processes also hold true for the silent observer. In particular, the satisfaction with the response and the credibility are important drivers to explain the cognitive processing. However, in contrast to the complainant, results indicate a higher importance of the outcome evaluation on comparison to the interactional judgement. The latter might affect beneficial strategies of companies to address both, the complaint and the observer, with one powerful answer. Limitations and Future Research

Although insightful, the present experiment has its limitations and raises questions for further research. First of all, the size of the sample that underlies the conclusions of the current study is relatively small. This circumstance may have prevented the discovery and/or substantiation of some of the effects that take place in the business practice besides those that were discussed above. It is a well-known statistical property that bigger sample sizes tend to increase the measure of significance of the results derived from them (e.g.,

3.4 Online Complaints in the Eye of the Beholder

125

Cohen, 2013). Though this property has to be treated with caution, it still can be argued that increasing the sample size has potentials to increase the knowledge resulting from this study. So, for example, a bigger sample size could allow us to make statements about the intentions of observers to spread negative WOM based on the observation of public complaint handling acts. Our current sample, however, doesn’t allow us to make such statements with the required confidence. With more data coming in, we hope to learn more details about the nuances of public complaint handling. Second, in this study (that we still consider to be a preliminary study) subjects were exposed to a rather simple scenario in that they had to form their opinion about a company based on an observation of a single complaint-handling act. It can be argued that such a scenario is too artificial, because in the real-world people tend to build their opinion based on several feedbacks and/or several feedback sources, rather than relying on a single public complaint-handling encounter. Future research should address this issue, i.e., it should simulate natural search for customer reports about a company’s performance in more complex situations in which the searchers face various opinions that differ both in direction and valance, some of them contradictory, and belong to different customers (like it can be seen, e.g., on amazon.com). Further, our study showed that offering monetary compensation produces stronger positive effects in observers than an apology does. However, offering money for every complaint is a costly strategy. Hence, to enable businesses to make optimal use of public offers of monetary compensation, further research should clarify when (i.e., in which situations) and how often such compensation should be offered as well as how high such compensations should be. Last but not least, the generalizability of this study’s results is hindered by the notion of cultural differences and legal liability, i.e., people may have concerns about being legally accountable for saying that a company has done something wrong without the proof and/or there might be differences that depend on the origin of the observers of complaint handling acts. In fact, this study only examined Germans, whose interpretations may deviate from those living in other countries influenced by other cultures. Yet, it can be argued that a study involving Germans depicts only the European way of thinking and doesn’t account for cultural differences in behavior of other nations (Wong, 2004).

126

3 Essays

3.4.6 Appendix Table 3.4-5. Overview Constructs Essay 4 Constructs (Cronbach’s Alpha)

Items

Sources

Attitude towards the company (α = .95)

is bad/ is good I don't like it / I like it is negative / is positive is unsatisfactory / is satisfactory do not please me / do please me

Adapted from Goldsmith et al. (2000)

Perceived trust towards the company (α = .90)

"online-deals.de" is trustworthy. I trust that "online-deals.de" would act to my advantage. "online-deals.de" would keep promises to me. I would trust in the information that "online- deals.de" gives me. The online retailer "onlinedeals.de" strives to be known for keeping its promises.

Adapted from Doney and Cannon (1997) and Jarvenpaa et al. (2000) and Koufaris and Hampton-Sosa (2004)

Liking of the company (α = .89)

"online-deals.de" very likeable. In the case that I would buy from the online retailer "onlinedeals.de", I would probably like it. I would expect most customers who shop at the online retailer "online-deals.de" to be satisfied with it. Online shoppers like me probably wouldn't like "online-deals.de".

Adapted from Esch/Geus (2005)

Riskattitude towards the company (α = .96)

a high purchase risk / a good bargain a possible disadvantage / a possible advantage a negative situation/ a positive situation a bad decision / a good decision

Adapted from Jarvenpaa et al. (2000)

Purchase intention (α = .86)

I think it is a good idea to buy the product "%produkt%" at "onlinedeals.de". I am positive about buying the product "%produkt%" at "onlinedeals.de". I think it is unwise to support the online retailer "online-deals.de" by buying from it.

Adapted from Laczniak/DeCarlo/Ramaswami (2001)

3.4 Online Complaints in the Eye of the Beholder

127

Positive WoM intention (α = .91)

I would very likely spread positive word-of-mouth about "onlinedeals.de". I would recommend "onlinedeals.de" to my friends. If my friends wanted to buy the product "%produkt%", I would tell them to try "online- deals.de".

Adapted from Maxham III/Netemeyer (2002a; 2002b)

Negative WoM intention (α = .88)

The probability is high that I will warn family and friends to not shop in this online store. If a problem like the one in the example happened to me, I would complain to my family and friends about this online store. If a problem like the one in the example happened to me, I would tell my friends and relatives that they should not shop in this online shop.

Adapted from Blodgett/Hill/Tax (1997)

Satisfaction with the answer (α = .91)

displease me/ please me I loathe/ I admire very dissatisfied me/ very satisfied me would not have helped me well / would have helped me well make me unhappy/ make me happy are worthless/ are valuable are frustrating / are gratifying

Adapted from Bansal/Irving/Taylor (2004) and Bansal/Taylor/Hames (2005)

Cause attribution (α = .74)

The customer has written the report to inform other customers exactly how good or bad the online retailer "online-deals.de" really is. I trust that the customer's report is based on his/her true experiences and feelings.

Adapted from Sen/Lermann (2007)

Usefulness of the answer (α = .91)

informative helpful valuable beneficial

Adapted from Gefen/Ridings (2005)

Perceived credibility (α = .94)

not credible / credible unreliable/ reliable dishonest / honest insincere / sincere

Adapted from Ohanian (1990)

128

3 Essays untrustworthy /trustworthy

Distributive justice (α = .93)

Even if the purchase caused problems for the customer, the efforts of "online- deals.de" to solve the problems led to a positive result. Considering the time and anger, the outcome that the customer received from "online-deals.de" was fair. Given the inconvenience caused by the problem, the outcome that the customer received from "online-deals.de" was fair. The outcome that the customer received in response to the customer service problem was more than fair.

Adapted from Maxham III/Netemeyer (2003)

Interactional justice (α = .88)

"online-deals.de" seemed to have great interest in the problem of the customer. "online-deals.de" understood exactly the problem. “online-deals.de" treated the customer roughly. "online-deals.de" tried very hard to solve the problem. All in all, the way in which "online-deals.de" responded to its customer's report was fair.

Adapted from Homburg/Fürst (2005)

Procedural justice (α = .77)

"online-deals.de" responded quickly to the customer's report. All in all, "online-deals.de"’s approach to the report was fair.

Adapted from Maxham III/Netemeyer (2003)

3.5 Differences and Similarities in Motivation for Offline and Online eSports Event Consumption 3.5.1 Introduction Roughly defined as “a form of sports where the primary aspects of the sport are facilitated by electronic systems”, eSports is already a key phenomenon of the modern digital area (Hamari and Sjöblom, 2017). Organized in leagues and ladders around different games of various genres, eSports is a very successful business venture and still growing year by year (Warman, 2017). Through streaming options on various platforms

3.5 Motivation for Offline and Online eSports Event Consumption

129

(e.g., twitch.tv or youtube.com), eSports can be consumed by users all over the world (Yu et al., 2018). These streams are extremely popular amongst eSports fans and are often consumed by millions of users (Warman, 2017). Additionally, eSports events, often hosted in big arenas and stadiums, allow thousands of eSports fans, who are willing to leave the purely digital environment of the internet, to consume eSports content in a completely new setting (Hallmann and Giel, 2017). Where users were previously constrained to consume eSports alone at home in front of their personal computer, they now fill arenas to watch their favorite team compete on stage (Hamari and Sjöblom, 2017). Hence, the digital barriers and limitations have vanished, and the overall experience has been enhanced to fulfill aspects of traditional events. The basic content, following two or more teams competing in a digital environment, remains the same for the event as well as the stream. Nevertheless, many of the surrounding factors do vary and might change the overall experience. In this context, uses and gratifications theory proposes that consumers decide on the basis of their needs (e.g., escape from reality) which form of media consumption is preferred and chosen (Hamari and Sjöblom, 2017; Katz, Blumler, et al., 1973; Katz, Haas, et al., 1973). Hence, this choice leads to a process which underlies the relationship between needs, the chosen consumption form and the gratifications obtained satisfying these needs (Palmgreen et al., 1985). Thus, the question arises how users are choosing a form of consumption, and what motivates them to attend the event on-site or follow the given stream online. Answering these questions is of importance especially for streaming service design (e.g., chat possibilities, custom camera views or other personalization options) and marketing potential e.g., for advertisers to align their advertising efforts to the consumer needs in the specific context. Research, thus far, has focused on a variety of aspects of general eSports consumptions but did not deal with the different forms of eSports consumption. Macey and Hamari (2017), Hallmann (2017) and Heere (2018) offered classification approaches of eSports with respect to other phenomena, as well as traditional sports, arguing for its general importance and overall social influence. The general consumption motivation of eSports has been assessed by Hamari and Sjöblom (2017), who developed a motivation scale that especially caters to eSports. Furthermore, Pizzo et al. (2017) as well as Donghun and Schoenstedt (2011), have analyzed the differences between sports and eSports consumption. Surprisingly, a comparison of the previously described two forms of eSports consumption is missing. Yet, literature regarding general sports consumption indicates possible differences between different forms of consumption which is predominantly indicated through differences in the motivation to follow the event (Hu et al., 2017; Seo and Green, 2008; Zhang and Byon, 2017). Thus, to get a more profound

130

3 Essays

view on differences between both consumption forms and therefore, be able to derive implications, our first research question reads as follows. RQ1: What differences can be observed in the motivation of onsite participants and online participants of eSports events? Moreover, studies have indicated that these differences might also impact important aspects of event success (Zhang and Byon, 2017). Relevant factors like satisfaction with the event and a corresponding attitude towards the event experience might, therefore, also be subject to the different forms of consumption. Thus, to get a more thorough view that goes beyond the differences in motivation regarding the event success, we strive to answer the following second research question in the context of eSports events: RQ2: What differences can be observed in the attitude towards the event and the satisfaction with the event between online and offline consumption form? By answering those two research questions, our study will widen knowledge on eSports consumption and assess the differences in off- and online consumption of eSports. To find explanatory ground for our research, we conducted a study at a league of legends event in Berlin. In the following, we will present the fundamentals of our research, the results of the study and derive implications for management and research. 3.5.2 Literature review and hypothesis development Conceptual Framework

Our conceptual model (see Figure 3.5-1) builds on motivations for eSports consumption in regard to the uses and gratifications, social-cognitive and general needs theory such as Maslow’s hierarchy of needs, two-factor theory and acquired-needs theory. Evalua#on and A:tudinal - Behavioral Outcome

Mo#va#on for eSports Consump#on (Needs) § § § § § § §

Social interac,on Vicarious achievement Acquisi,on of knowledge Skills of the players/athletes Aesthe,cs Escape Drama

Form
of Consump#on offline (analogues) vs. online (digital)

§ Hedonic ADtude towards the Event § U,litarian ADtude towards the Event § Sa,sfac,on with the Event § Inten,on to par,pcate offline vs. online



Figure 3.5-1. Conceptual Model

In general, we believe that choosing a specific form of consumption of an eSports event is driven by the actual motivation for eSports consumption, because consumers have

3.5 Motivation for Offline and Online eSports Event Consumption

131

specific needs, that they expect to be fulfilled via a specific media form. Interactions with the environment will lead to an evaluation of consumers, if their needs are satisfied by the consumption form and consequently, a positive or negative confirmation affects future media consumption behavior. Motivation as needs for sports consumption In general, uses and gratifications theory is one of the most employed frameworks to understand media use and a related digital and nondigital consumption (Hamari et al., 2018; Salwen et al., 2004). Specifically, this framework helps to understand consumer motives for accessing and using a form of media consumption (Katz, Haas, et al., 1973). Consumers decide on the basis of interests (e.g., context) and needs (e.g., escape from reality or entertainment) which media offer is consumed (Hamari and Sjöblom, 2017). The decision for a specific medium (i.e., in this context the consumption form for eSports) depends thus on the user’s motivations and expectations that their needs are satisfied by the media offered and therefore, a congruency between needs and an expected fulfillment (Severin and Tankard, 2013). In general, motivations are understood as a predisposition that drives human behavior with the ultimate goal to fulfill basic to more complex needs (Maslow, 1943; McDonald et al., 2002). Li and Petrick (2005) emphasize that the link between socio-psychological needs of consumers and their resulting motivation to participate in specific events has created a significant basis for studies on event motivation research (Crompton, 2003). Based on two-factor and acquired-needs theory literature supposes that the motivation of consumers participating at events is shaped by their needs and the expectations to fulfil these sociopsychological needs (Getz, 1991; Li and Petrick, 2005). Extending these assumptions by the social-cognitive theory, literature suggests that enactive learning renders this evaluation of fulfillment and therefore, affects related behavior (Larose et al., 2001). Enactive learning describes how humans learn from experience and consequently, how interactions with the environment (e.g., consumption form) influence media presence by (continuously) reforming expectations and evaluations of the likely outcomes of future media consumption and the related behavior (Bandura, 1986; Larose et al., 2001). It is therefore a process that describes the relationship between gratifications sought, media behavior, and gratifications obtained (Palmgreen et al., 1985). Consequently, the expectations of satisfaction of needs lead to the choice of a specific consumption form and consequently, the exposure to this specific media environment determines future behavior due to consumers’ evaluations (Kaye et al., 2018; Larose et al., 2001; Sundar and Limperos, 2013). Hamari et al. (2018) emphasize that the uses and gratifications framework has been used by research from various domains to investigate motives associated with different contexts and especially in the context of video games, such as

132

3 Essays

video games in general (Chen and Leung, 2016; Kim and Ross, 2006; Merhi, 2016), online video game streaming (Hilvert-Bruce et al., 2018; Sjöblom and Hamari, 2017) and eSports in particular (Hamari and Sjöblom, 2017). Moreover, differences in offline and online consumption motivation (e.g., news consumption form) have been assessed by taking the uses and gratifications concept into account (Salwen et al., 2004). General motivations to attend events have been studied for several years and researchers strived for a more profound knowledge to understand the different groups of attendees to better meet their needs. To do so, research focused on developing scales that include the relevant motivation dimensions. For instance, Uysal, Gahan and Martin (1993) were among the first to develop a scale that dealt with the different dimensions of event attendees’ motivations. Specific elements of spectator motivation were identified through additional research and coherent measurements derived (McDonald et al., 2002; Wann, 1995). Versions of the scale were adapted to various settings and tested at numerous events, showing its general usefulness to describe the event participation motivation. Moreover, Scholars utilized different dimensions of motivation to, e.g., segment different groups of visitors (Backman et al., 1995; Formica and Uysal, 1995). Trail and James (2001), subsequently, build on the known scales and coherent dimensions of motivation to develop the motivation scale for sport consumption (MSSC), that was designed to measure the general motivation for sport consumption (e.g., basketball or soccer), independent from a specific event. Their work, subsequently, was verified through numerous studies that assessed broader aspects of sport consumption and related behavior, e.g., the differences of male and female sport fans (Byon et al., 2010; Ridinguer and James, 2002) and, thus, demonstrating the range of possible applications for the scale to assess motivation. Recently, Hamari and Sjöblom (2017), build on the MSSC to develop a tool that was applicable to eSports. Consequently, they are able to derive insights about the comparability of eSports and traditional sports and extend the knowledge about eSports consumption. Nevertheless, the form of eSports consumption (i.e., digital or analogue) is not subject to their assessment. Yet, the motivation for eSports consumption and the related decision for a media form is particularly interesting, because eSports roots in digital environments, whereas classical sports (e.g., soccer or basketball) originated from an analogue world of consumption (Hamari and Sjöblom, 2017; Zhang and Byon, 2017). Thus, the online consumption form of eSports existed before offline consumption forms (i.e., non-purely digital events) were introduced (Scholz, 2019). Consequently, we build on the work of Hamari and Sjöblom (2017) and assess differences of the eSports consumption forms based on the adapted version of the MSSC. This scale consists of the dimensions taken from the original MSSC by Trail and James (2001), however, adapted to the eSports context:

3.5 Motivation for Offline and Online eSports Event Consumption

133

n The dimension social interaction refers to the notion that people consuming eSports seek contact with like-minded individuals (Hamari and Sjöblom, 2017). Generally, the possible association with an (unknown) group or individuals and socializing opportunities with friends, business partners or other acquaintances has been found to be an important driver for event attendance and sport consumption alike (Crompton and McKay, 1997; McDonald et al., 2002). n Experiencing vicarious achievements through, e.g., successful games by a player or team is the second motivation dimension (Hamari and Sjöblom, 2017). Witnessing how the favorite player wins an important game or successfully executes an important play evokes positive feelings for the spectating fan (Trail and James, 2001). The desire, to experience these feelings and be part of the moment, potentially drives fans to follow events. n Especially for attendees that are also playing the given game at home, the possibility to acquire knowledge might be of importance as a motivation to consume eSports. By following the live gameplay of a professional eSports game, where the most skilled players are competing with one another, spectators can learn new strategies or tactics that enable them to strengthen their own gaming performance (Hamari and Sjöblom, 2017). n Spectators might also be interested in experiencing the skillset of professional players and enjoy the performance of the high level of gameplay (Hamari and Sjöblom, 2017). The importance of this dimension can especially be found for active players, as they can truly appreciate and evaluate specific skills by comparing the gameplay with their own performance in the game (McDonald et al., 2002). n Likewise connected to the game itself is the perceived aesthetics of the sport. This dimension relates to the grace and beauty that is coherent in the game (Hamari and Sjöblom, 2017). Especially for sporting activities like ice skating or gymnastics, this aspect can be very important to draw a crowd to the event (Milne and McDonald, 1999). Relating to eSports, the maneuver of in-game characters or the beauty of the displayed scenery might evoke similar feelings for possible attendees, motivating them to attend or follow an eSports event (Hamari and Sjöblom, 2017). n A more general dimension of motivation to attend or follow an eSports game can be found in the desire to escape from daily routines. The game, the event and the, for most, unusual occasion of experiencing it live deals as a distraction from regular life (Gantz and Wenner, 1995). Reducing stress by escaping to another environment and switch into another mindset as long as the event lasts has been

134

3 Essays

found to be an important reason for actively participating as well as spectating an event (Milne and McDonald, 1999). n Lastly, the coherent drama of undecided games and close competition amongst rivals also motivates visitors to attend an event (Hamari and Sjöblom, 2017). Enjoying the atmosphere of important games (e.g., championship finals) evokes emotional reactions that attendees were found interested in experiencing (Trail and James, 2001). These dimensions (social interaction, vicarious achievements, acquisition of knowledge, skillset of professional players, aesthetics, escape and drama) build the MSSC that was already utilized to measure general eSports consumption (Hamari and Sjöblom, 2017; Katz, Haas, et al., 1973). However, based on previous research, we propose that the actual environment of consumption impacts the overall experience (Seo, 2013) so that different forms of participation (i.e., pure digital online vs. offline at the event venue) might shape different expectations regarding which form of consumption satisfies the needs best. Hence, in the following we will present reasoning for potential influences of these motivational factors on the choice of consumption form and derive respective hypotheses. Hypothesis development One key element of events is related to interaction of people with one another. Often, groups of friends or family attend an event together and use the provided content as a platform for their social interaction with each other (Kerr and May, 2011; Pons et al., 2006). This is something that also holds true for eSports in general (Hamari and Sjöblom, 2017). Nowadays, technology allows for interaction with other users in virtual places. Streaming platforms, e.g., twitch.tv, have integrated features that allow contact with other individuals while consuming an eSports stream (Bründl et al., 2017; Scheibe et al., 2016). Therefore, the basic possibility of interaction is provided in both consumption scenarios. However, researchers have argued that the virtual interaction with peers or family is often seen as a substitute for real life interaction (Seo and Green, 2008). Users of streams could certainly be interested in using interaction features of provided platforms, but the social connection is much more relatable to a real-life interaction provided by event. Hence, we argue that there will be differences in offline and online eSports consumption in the social dimensions of motivation. H1: Social motivation to participate in an event will be significantly higher for offline than for online participants. Next to the socialization with other visitors or users of an eSports event, the perceived social connection to the players is also an important motivational factor for (e)sport consumption. Experiencing a victorious achievement and celebrating the success of a

3.5 Motivation for Offline and Online eSports Event Consumption

135

favorite player is considered to be an important motivational factor of all sport spectators (Dale et al., 2005; Fink et al., 2002; Gantz and Wenner, 1995). When comparing the two consumption possibilities of eSports, one can argue that the offline consumption allows for a stronger connection with other fans and spectators, while the online consumption enhances the perceived connection to victorious players. eSports has been an online phenomena and most active players are still using websites, social media and other virtual communities to present themselves (Hamari and Sjöblom, 2017; Seo, 2013). Events are a sort of exception to these normal representations, that are hosted irregularly and sometimes far away from specific fans (Seo and Jung, 2016). Nonetheless, fans of specific players will be able to follow their favorite team or player online. In a successful game, their fan-based perceived connection will, as it does in most sports, lead to a perceived level of combined success, where the victory will in turn be perceived as a personal achievement (Hamari and Sjöblom, 2017; Wann, 1995; Wann and Branscombe, 1993). Attendees of the event will certainly not be free of this motivational dimension, but the perceived achievement of online users will be, based on this reasoning, significantly higher: H2: Achievement motivation to participate in an event will be significantly higher for online than for offline participants. Moreover, gaining knowledge has been shown to be a relevant factor for sport consumption, e.g., learning about the players or teams and sharing this information in conversations about sports or to improve own skills in the respective sport (Hamari and Sjöblom, 2017). Attending any form of sporting event generally offers different forms of knowledge acquisition. One aspect can be found in the possibility of attendees to inform themselves about the venue, players and teams (Gantz and Wenner, 1995; Wenner, 2013). Furthermore, information about the sport in general, e.g. tactics or play styles, can be obtained by attendees (Karp and Yoels, 1990). Users following the stream, or people attending the event, are also very likely to play this game themselves. Experiencing other (professional) players playing the game offers the possibility to extend their own degree of knowledge about the game and possible strategies and tactics. Both dimensions are, therefore, expected to influence any form of eSports consumption. However, Hamilton et al. (2014) have discussed the importance of knowledge sharing in online media consumption settings and stated that new streaming platforms offer enormous potential to exchange expertise about the issue. Yet, differences in acquisition of knowledge should be a key difference between the two consumption forms, leading to different ways of game portrayal. People at the event will most likely not be as close to the action as streaming users. Building on additional features of twitch and similar

136

3 Essays

websites, users are enabled to follow the action intensively and learn about the game, the players’ tactics and strategies. Therefore, we assume that the motivation to obtain knowledge will be significantly higher for stream users: H3: Acquisition of knowledge motivation to participate in an event will be significantly higher for online than for offline participants. Opportunities to experience the skillset of players are provided for online and offline consumption forms (Pizzo et al., 2017). However, the provided features of streaming platforms exceed the event attendees’ point of view in the arena, e.g., to improve own skills (Hamari and Sjöblom, 2017). Where event attendees are, by design, forced to follow a broader overview of the game and the related action, stream users are enabled to follow the game closely and appreciate the skillset of players (van Hilvoorde and Pot, 2016; Weiss and Schiele, 2013). The implemented platforms even allow users to switch between different viewpoints, enabling them to exclusively follow individual players and obtain a better understanding of their tactics. For instance, stream viewers can actively switch and choose between different game perspectives such as overview cameras, but especially third-person and first-person views (e.g., seeing directly what the athlete sees), which are unique to eSports and which affect remarkably how the consumers can follow a match. Here, some event broadcasts even offer complete freedom to move within the in-game environment (e.g., Defense of the Ancients II), which is superior to traditional broadcasting means such as soccer or basketball. Hence, these digital environments offer static streaming broadcasts, where multiple camera angles are chosen by the moderator. However, they also offer active environments where the observer can directly choose the camera perspective. Consequently, the online consumption of an event offers more comprehensive opportunities to follow individual players and performances in comparison to offline consumption. Therefore, the motivation to appreciate the skillset of the involved players will be significantly higher for online participants. H4: The motivation to experience the skillset of professional players will be significantly higher for online than for offline participants. Another aspect of the game, that motivates potential spectators, is the aesthetic demonstration of players. Relating to the elegance or excellence of the sport, this motivational trait is especially influential in very visual sports that allow spectators to observe a detailed form of sport (Hamari and Sjöblom, 2017). Therefore, sports that allow, or even generally include the judgement of strong visual elements, e.g. gymnastics, are commonly considered to attract viewers with a strong aesthetic motivation (Fink et al., 2002; Trail and James, 2001; Wann, 1995). Here, Hamari and Sjöblom (2017) found that eSports consumption was negatively influenced by the aesthetic motivation of users. They argued that the basic link between this motivational

3.5 Motivation for Offline and Online eSports Event Consumption

137

dimension and the eSports consumption was very well given, but that the form of utilization as well as the game genre in question would play an important role. In deference to traditional forms of sport, most games played feature long and intensive battles. Therefore, the possibility to enjoy and observe specifics of the players’ skillset are rather limited. Other forms of sports, e.g., gymnastics or golf, do offer a relaxed setting that allows spectators to observe the performance of a single athlete while most games played in eSports are based on interaction of two or more teams with almost no break. Because of a continuous gameplay and interaction of players, the opportunities of spectators to focus on a single player’s performance is limited. Nevertheless, a general possibility of enjoying an aesthetic performance is certainly given in both forms of consumption and build on the discussed advantages of the existing platforms. One example of these eSports aesthetics might be the players’ performance with the mouse and keyboard, i.e., the so-called (and often depicted) actions per minute. However, these actions need close ups of the players’ hands, which are more usually broadcast within streams (e.g., by picture in picture), whereas the offline consumption, i.e. the big screen at the event, mostly focuses on the actual gameplay. Thus, we argue that the online participants will show a significantly higher aesthetic motivation based on the consumption possibilities: H5: Aesthetics motivation to participate in an event will be significantly higher for online than for offline participants. Moreover, based on previous research into sports event consumption, we argue that an escape from daily routines is another dimension of motivation. The content observed might be used as a distraction from problems and issues that might bother the individual (Gantz and Wenner, 1995). Sports consumption in general, and eSports consumption specifically, have been shown to cater for this dimension of motivation (Hamari and Sjöblom, 2017; Trail and James, 2001). Offline and online consumption of eSports events should, therefore, be able to provide possibilities for escapism to users and attendees alike. However, recent research shows that the actual environment of consumption impacts the overall experience (Seo and Green, 2008; Seo, 2013). Thus, consuming a stream at home might be less effective in creating an escape perception, because the environment (e.g., in front of a PC or television) is still like other daily experiences. On the other hand, visiting an event on-site (e.g., arena) offers new and as yet unknown impressions and thus, should be sought by consumers with a more distinct desire for escape. Based on Kahneman’s (1973) assumptions regarding attention and efforts, we propose that new impressions lead to a need of higher capacities to process the surrounding information and thus, reduce the probability to think about mundane things. Hence, we hypothesize:

138

3 Essays

H6: Escape motivation to participate in an event will be significantly higher for offline than for online participants. Moreover, the drama of a match is another dimension of motivation and might be very similar to the previous dimension in the case of the impact of an offline event. Drama refers to the uncertain outcome of games. A close game that offers a lot of excitement to viewers is a key element of (e)sport consumption (Pizzo et al., 2017; Pons et al., 2006; Trail and James, 2001), since the content provided offline and online is identical and allows both groups to experience the game and its outcome. However, drama might be interpreted as multidimensional and thus, should be affected by more influential factors than just the outcome of a match. For instance, the overall atmosphere in an arena with thousands of spectators following an extremely thrilling game situation should intensify the perceived drama. Similar results can be observed, for example, in research into basketball or other forms of sport consumption (Chen et al., 2013; Zhang and Byon, 2017). Thus, eSports enthusiasts with a more distinct need for drama, should seek offline event participation: H7: Drama motivation to participate in an event will be significantly higher for offline than for online participants. Attitude and satisfaction The attitude towards the event has been identified as a key factor to explain event-related behavior and measure the overall success of events (Martensen and Gronholdt, 2008). Especially in regard of sponsoring effects, the attitude towards the event and the related brand have shown significant influence (Carrillat et al., 2005; Ruth and Simonin, 2003). Therefore, eSports events and offerings should be keen on understanding the influential factors of attitude towards the event and how it is related to the form of consumption (Macey and Hamari, 2017; Seo, 2013). Hence, we argue that the attitude towards an eSports event is also an important factor to be assessed when analyzing the different consumption forms. Gursoy et al. (2006) introduced the concept of two dimensional attitude towards an event. With the distinction of utilitarian and hedonic aspects, they argue, the different factors of event consumption can better be described this way (Gursoy et al., 2006). Similar approaches have also been brought forward in digital environments where Salehan et al. (2017) have found reasoning that both dimensions are also relevant to explain the behavior of users in social networking services. Hedonic attitude of event consumption relates to aspects of enjoyment and perceived fun yielded through the event (Gursoy et al., 2006). These aspects may be perceived differently from individual to individual, but a general understanding that this dimension plays a vital part in explaining attitude towards an event is assumed (Gursoy et al., 2006). In digital environments, hedonic attitude has been connected to self-enhancing

3.5 Motivation for Offline and Online eSports Event Consumption

139

and joyful experiences, that are also perceived individually (van der Heijden, 2004). In particular, research into social networking sites has addressed this issue and concluded that the social features (e.g., connecting with other users) are very relevant to explain the perceived enjoyment of involved users (Salehan et al., 2017). In regard to electronic gaming, research has also identified social interaction to play a vital role in explaining the hedonic attitude of users (Salehan et al., 2017). As previously stated, the environmental setting of offline consumption will enhance the perceived connection of attendees. Therefore, we argue that the overall attitude towards the event will be significantly higher for offline participants: H8: Hedonic attitude towards the event will be significantly higher for offline than for online participants. Utilitarian factors relate to the possibility of event attendees or stream users to utilize the experience to their advantage (Gursoy et al., 2006). In digital environments, e.g., social networking sites, users tend to advance their career by connecting with possible employers online, or sharing and gathering job-related information (Salehan et al., 2017). Furthermore, users tend to visit websites as a source of knowledge that enhances their private or professional life (Ardichvili et al., 2003). Similar effects can be expected in regard of streaming options of eSports events. Websites are often conceived as a tool that enables users to enhance their personal or private life. Therefore, users’ utilitarian attitude towards the event is likely to be higher for online participants as their focus of consumption is likely to be strongly connected to factors such as knowledge gain and aesthetic appreciation to enhance their own skillset: H9: Utilitarian attitude towards the event will be significantly higher for online than for offline participants. Moreover, event related satisfaction has been considered to be connected to the game and the service satisfaction (Kim et al., 2016; Yoshida and James, 2010). Game-related satisfaction would be tangible in both consumption forms, while service satisfaction would certainly be conceived differently in both settings. Yoshida and James (2010) argue that the atmosphere is a strong indicator for overall satisfaction. Within online environments the satisfaction might, therefore, be related to the community and their connection with one another, but real perceived atmosphere is only conceivable within offline forms of consumption (Steinmann et al., 2015). Therefore, we postulate: H10: Satisfaction with the event will be significantly higher for offline than for online participants.

140

3 Essays

Behavior Forthcoming event success is highly related to positive behavioral intention of visitors. Through their revisiting intention, they can positively influence the long-term success of events. Kim et al. (2016) found that revisiting intentions are strongly related to the experiences made while attending the event. Therefore, we assume that either form of consumption will lead to visiting intentions of the participants. Furthermore, we argue that event attendees on site are more likely to show an intention to visit the event on-site again, while online consumers might tend to higher interest in watching another streamed version of an eSports event. Nevertheless, it should be noted that consumers will tend to participate in forthcoming events either way if the needs are satisfied (e.g., cross media usage), however, consumers will tend to show a higher level of loyalty to participate in an environment that satisfies best their consumption needs (Larose et al., 2001). H11: The intention to visit an event on site will be significantly higher for offline than for online participants. H12: The intention to consume a stream of the event will be significantly higher for online than for offline participants. 3.5.3 The empirical study Measures and procedures To test our hypotheses, we prepared a questionnaire for the EU LCS Event in Berlin in early 2018. Riot Games, organizer of the event, offered exclusive live coverage of the event through lolsports, youtube and twitch.tv. However, the actual content (i.e. the video stream) was similar on all three websites. The same applies, for instance, to the interaction possibilities (e.g. chat), so that these three websites can be classified as highly comparable. The coverage included commentated gaming content as well as shots from inside the event venue. This is a standard form of eSport online event coverage and provides the desired background for our study. In accordance with the language spoken at the event and in the online stream, the survey was conducted in English. Hence, everyone following the stream was able to take part in our survey. By utilizing international, game-related message boards (e.g. reddit and twitter) to reach online participants, we furthermore ensured that a representative, international sample could be drawn. On-site participants were randomly approached with a similar paper and pencil version of the questionnaire. At the beginning of the questionnaire, participants were asked what form of consumption they had chosen, i.e., on-site or online consumption, to ensure that participants could be unequivocally assigned to either one of the two groups. Moreover,

3.5 Motivation for Offline and Online eSports Event Consumption

141

participants were clearly instructed to only access the previously selected event form to guarantee a high degree of discriminatory power. In addition to demographics, we used measures that related to the postulated hypotheses. The motivational dimensions were operationalized in accordance with Hamari and Sjöblom (2017) and the MSSC of Trail and James (2001) in the context of eSports. Both dimensions of attitude were measured with five items each, taken from Gursoy et al. (2006). Satisfaction with the event was adapted from Voss et al. (1998). Intentions were measured with one item taken from Wakefield (1995). The measurement was performed using well established multi-item scales with a seven-point Likert scale and all reflective constructs satisfy the Cronbach’s Alpha threshold of > 0.70. For a detailed overview in regard to the measures used in this essay, see appendix Table 3.5-2. The final sample, both on-site and online, consisted of N = 637 participants (81.7 % male, mean age M = 21.40, standard deviation SD = 5.59). Of these, online viewers: n = 482 respondents (86.9% male, age M = 21.01, SD = 4.65), and on-site participants: n = 155 respondents (65.1% male, age M = 22.73, SD = 7.79). Results and discussion To verify our hypotheses, we used multiple t-tests with offline (on-site participation) and online consumption via stream as independent variable. Table 3.5-1 shows the results of our analysis. The reason for choosing t-tests is that research has shown t-tests to be robust against violation of statistical requirements (e.g., different group sizes or non-normal distribution) (Glass et al., 1972; Sawilowsky and Blair, 1992). In addition, as we are comparing two groups, i.e., offline versus online consumption, using t-tests seems appropriate. Results show significant differences with regard to almost all variables under review. Most hypotheses can be validated via the derived results. Firstly, regarding motivation, we mostly observe the expected differences. Here, the social dimension is more pronounced in case of offline events. This dimension can, therefore, be considered as more relevant in an offline environment and seems to be more likely to be supported by a traditional form of event consumption, i.e., meeting friends and family at an event. However, social interaction cannot be described as the primary driver of consumption, as it tends to be less important in comparison to the remaining dimensions. Thus, the result is also interesting in terms of knowledge gain and the observation of player's skills. Both dimensions are more distinctive of stream consumption. The latter might be explained by the details within the digital stream, i.e. player close ups and direct screen capturing directly on the screen at home, which enables the consumers to follow the matches in detail. In comparison, offline participants, who can “only” follow the match on a huge canvas, do not get that level of

142

3 Essays

detail. This assumption might also be supported by taking the results of the aesthetic dimension into consideration. Nonetheless, it should be mentioned that while all the dimensions differ significantly, the size of the effect is rather small. Table 3.5-1. Hypothesis testing Dependent Variable

Mean (SD)

t-Value (p-Value)

Hypothesis

Social

Online: 4.32 (1.70) Offline: 4.75 (1.76)

T(633) = 2.721 (p = .007)

H1 ü

Achievement

Online: 5.08 (1.55) Offline: 4.78 (1.77)

T(633) = 2.086 (p = .037)

H2 ü

Gain Knowledge

Online: 5.95 (1.05) Offline: 5.46 (1.30)

T(633) = 4.761 (p < .001)

H3 ü

(Physical) Skills

Online: 6.52 (0.75) Offline: 6.19 (1.14)

T(633) = 4.028 (p < .001)

H4 ü

Aesthetics

Online: 5.56 (1.29) Offline: 5.08 (1.49)

T(633) = 3.873 (p < .001)

H5 ü

Escape

Online: 4.57 (1.47) Offline: 4.70 (1.67)

T(633) = .923 (p = .357)

H6 û

Drama

Online: 6.16 (1.03) Offline: 6.04 (1.27)

T(633) = 1.171 (p = .242)

H7 û

Hedonic Attitude

Online: 6.19 (0.96) Offline: 6.18 (1.18)

T(633) = 0.012 (p = .990)

H8 û

Utilitarian Attitude

Online: 5.26 (1.09) Offline: 5.31 (1.08)

T(633) = .512 (p = .609)

H9 û

Satisfaction with the Event

Online: 5.65 (1.10) Offline: 6.00 (1.21)

T(633) = 3.364 (p = .001)

H10 ü

Attend Offline

Online: 2.16 (1.69) Offline: 4.41 (2.06)

T(633) = 13.59 (p < .001)

H11 ü

Attend Online

Online: 6.21 (1.18) Offline: 5.69 (1.86)

T(633) = 4.007 (p < .001)

H12 ü

Motivation to attend

Attitude and Satisfaction

Behavior

1 = totally disagree / negative evaluation, 7 = totally agree / positive evaluation, insignificant results are italic

3.5 Motivation for Offline and Online eSports Event Consumption

143

Surprisingly, we do not find any effect regarding the attitude dimensions towards the event. Generally, the data shows that the event, in both consumption forms, is considered to yield hedonic as well as utilitarian features. For hedonic, M = 6.19 and 6.18 and for utilitarian, M = 5.26 and 5.31 (online vs. offline, respectively), the overall values for hedonic attitude are more pronounced in comparison to the derived values for utilitarian attitude. eSports is, first of all, based on a game that obviously is played for the enjoyment it yields. Nevertheless, the high value for utilitarian attitude demonstrates that eSports also offers a lot of useful aspects to its fans. In accordance to the data received for the motivational subscales that relate to utilitarian aspects (e.g., knowledge gains), the analysis generally indicated the importance of these factors. Prior research indicated that most events and products can very well cater to both dimensions of attitude, and our research supports those claims (Batra and Ahtola, 1991; Gursoy et al., 2006). Nonetheless, the proposed differences between the two consumption methods cannot be observed, leading to the assumption that the overall attitude towards the event manifests on a different level and is not directly determined by the chosen form of consumption. As hypothesized, the satisfaction with the event does in fact differ among the two forms of consumption. Generally, the perceived satisfaction of the participants is relatively high in both groups, indicating a positive reaction to the event. Building on the argument and research of Yoshida and James (2010), we argued that the atmosphere and surrounding factors (e.g., form of broadcast in the arena) have a more positive influence on the level of satisfaction than the surrounding factors of online consumption. Here, satisfaction in regard to digital experiences is generally considered to be highly dependent on the surrounding factors that users create for themselves (Anderson and Srinivasan, 2003). Therefore, the possibilities for event organizers concerning streaming options are limited to the utilized platform. Everything else is ultimately left to the users’ own efforts to enhance the experience. On the other hand, the factors influencing event satisfaction for visitors on-site are much more tangible for the organizers (Yoshida and James, 2010). Event-related research has indicated numerous factors that, directly or indirectly, influence the perceived satisfaction with the event, all of which can and should be addressed by the event organizers (Chen et al., 2013; Kim et al., 2016). Regarding the behavior of the participants, we see differences in both variables examined. While both groups intend to watch another event online, participation in an offline event reveals a different result. Offline participants would tend to participate again, whereas the results show that online participants would continue to stick to the stream only. Online streaming has become easily accessible for almost everyone with a fast enough internet connection (Bründl et al., 2017; Chen and Lin, 2018). Hence, there

144

3 Essays

are few obstacles to witness another eSports event online. Fans of the game and the event series will always be interested in witnessing another event. Streaming certainly seems to be perceived as the more convenient option. However, on-site event participation does offer additional features of personal connection and atmosphere, but the main consumption method for most attendees and followers seems to remain within the digital environment. 3.5.4 Conclusion Summary of findings Overall, we were able to identify differences as well as similarities between both forms of consumption. The differences in motivation to consume provide further proof of the strengths and weaknesses of eSports events and streams. Keen observers and fans of the game, who are interested in playing themselves, seem to favor the streaming option, just to be able to examine the action closely. Events offer more emotional fans a great outlet for social interaction. Nevertheless, similarities in attitude and some motivation dimension show that the general perception of the event does not differ significantly between the two groups. Attending events on-site and following a stream online, based on our data, cannot be considered a substitute for each other. Each consumption method offers advantages, based on slightly differently motivated visitors and consumers. Given that even important outcome variables (e.g., satisfaction) differ for both forms of consumption, it is important to address the advantages of each form and cater to their strengths. These lead to interesting research questions and implications for managers. Implications and limitations Building on recent research results in the fields of eSports, event-marketing and online environment, our research helps to widen the existing literature in this new field of interest. eSports is a global and rising phenomenon with unique features that need to be addressed. The derived differences of both forms of consumption indicate that the motivational dimensions related to performance and the game itself were significantly higher for online participants. The question arises as to what characterizes these participants. Similar research into Chinese table tennis matches, for example, has shown that online participants demonstrate stronger feelings of fanship with players (Zhang and Byon, 2017). Further examination of the participants should, therefore, be addressed in further research. Moreover, our study was conducted at a League of Legends event in Berlin and online. Due to this setting, a number of limitations arise. eSports includes numerous games of different genres (Seo and Jung, 2016). Therefore, it is very likely that the derived results

3.5 Motivation for Offline and Online eSports Event Consumption

145

differ when assessing a different game from a different genre (e.g., tactical shooter like Overwatch or Counter-Strike). The event type has been found to play a vital role in traditional event-related research, and similar aspects could be connected to the game played when dealing with eSports (Crompton and McKay, 1997; Kim et al., 2016). In this context it should also be mentioned that geographical and economic limitations might affect the present results (Backman et al., 1995; Chung and Woo, 2011). The latter could explain why, on the one hand, we find significant differences between offline and online eSports consumption motivation, but on the other only observe relatively small effect sizes. Here, event observers who answered the questionnaire regarding online participation might have an eSport motivation, which would lead to the conclusion that those gamers prefer on-site consumption. However, due to considerable economic expense (for example the cost of traveling from their own country, potentially a long way from the event), simply cannot participate offline. Hence, further research could address this issue and investigate the impact of an offline consumption willingness in context of a “forced” online participation, i.e. stream. Our sample portrays a common issue regarding eSports research. Most of the players and followers, thus far, are male (Hartmann and Klimmt, 2006). Although the issue of gender has been addressed by eSports-related research (Gray et al., 2018; Kaye et al., 2018), the male dominance of participation limits the possibilities to fully assess this influential factor, even though our sample represents the underlying population’s gender distribution and provides sufficient explanatory power for the conducted study. Although literature argues for a connection between motivational factors and attitude towards an event, our results show that the effect of the consumption form is only given in the motivational factors. Event-related research has, thus far, only assessed the motivational factors of event visitation (Backman et al., 1995; Kulczynski et al., 2016; Lee, Lee and Wicks, 2004) or argued for the value of attitude to explain sponsorship effectiveness and other phenomena (Carrillat et al., 2005; Lee et al., 1997; Martensen and Gronholdt, 2008; Martensen et al., 2007). Future research endeavors should try to connect these issues and learn about the interplay of these two constructs. Human behavior in social live streaming services has been assessed through several studies, addressing factors such as platform representation, identification and interaction with streamers, and consumer expectations (Bründl et al., 2017; Oyedele and Simpson, 2018; Scheibe et al., 2016). Assumptions derived from these studies build on the usage of twitch and similar platforms to follow an individual or a company. The special aspects of event consumption (i.e. eSports event consumption) has not yet been addressed.

146

3 Essays

Social factors were among the few aspects of motivation that demonstrated stronger values for offline participation. Modern streaming platforms offer numerous options to communicate with other users (i.e. through direct message or chat), but these options are not yet fully capable of replacing real life experiences (Bründl et al., 2017; HilvertBruce et al., 2018; Scheibe et al., 2016). Lim et al. (2012) evaluated the influence of the perceived psychological distance of streaming users, and their research indicates that there are a few things that platform designers could implement to strengthen the perceived tie of users. Accordingly, eSports managers could possibly enhance the social experience of users when streaming an event. Through group offerings, special chat rooms, and more interactive features the perceived social connection of users could be enhanced. Another aspect of possible social interaction could be seen in the connection between players and their fans. Our data also indicates that players, their skillset, and the possibility of knowledge gain are advantageous features of online consumption. These aspects could also be enhanced by a more personal connection between players and the audience. Seeing that these aspects seem to be of importance to eSports fans, additional offerings that allow for a more personal and intense interaction of attendees, users, and the players should lead to positive reactions of fans (Seo and Jung, 2016; Weiss and Schiele, 2013). Research has indicated that stream followers are often interested in a personal connection with the streamer and that the perceived connection can also enhance positively related features (e.g., trust or fanship) (Friedländer, 2017; Hu et al., 2017; Yu et al., 2018). Due to the digital origins of eSports and its connection to the streaming community, the personality of players should be considered as an asset that needs to be addressed more strongly by event organizers. 3.5.5 Appendix Table 3.5-2. Overview Constructs Essay 5 Constructs (Cronbach’s Alpha) Motivation Scale for Sports Consumption:

Achievement (α = .87) Aesthetics (α = .90)

Items

Sources

I feel proud when my preferred player does well. I feel a personal sense of achievement when my preferred player does well. I feel like I have won when my preferred player wins. I appreciate the beauty inherent in video games. I enjoy the natural beauty in gaming I enjoy the gracefulness associated with gaming.

Adapted from Trail/James (2001) and Hamari/ Sjöblom (2017)

3.5 Motivation for Offline and Online eSports Event Consumption Escape (α = .73)

Drama (α = .76)

Player Skills (α = .86)

Social (α = .87)

Knowledge (α = .82)

Utilitarian attitude towards the event (α = .85) Hedonic attitude towards the event (α = .89) Satisfaction with the event (α = .91) Intention to participate

Watching eSports is a change of pace from what I regularly do. Watching eSports provide an escape for me from my day-to-day routine. Watching eSports provides a diversion from “life’s little problems” for me. I enjoy the drama of a close match. I enjoy it when the outcome is not decided until the very end. I prefer watching a close game rather than a one-sided game. The skills of the players are something I appreciate. I enjoy a skillful performance by the players. I enjoy watching a well-executed gaming. I enjoy interacting with other spectators at the game. I enjoy talking with others at the match. I enjoy socializing with people sitting near me while I watch the match. I increase my understanding of the strategies by watching matches. I increase my knowledge about a particular game when I watch it. I can learn about the technical aspects of a game by watching it. Necessary / Unnecessary Effective / Ineffective Functional / not Functional Practical / Impractical Helpful / Unhelpful Dull / Exciting not Delightful / Delightful not Fun / Fun not Thrilling / Thrilling Boring / Interesting I am satisfied with the Event. I am happy with the Event. I am delighted with the Event. In the future, how often will you attend EU LCS events? In the future, how often will you watch EU LCS events online? very infrequent / very frequent

147

Adapted from Gursoy et al. (2006)

Adapted from Gursoy et al. (2006)

Adapted from Voss et al. (1998) Adapted from Wakefield (1995)

148

3 Essays

3.6 The Need for a Community: The Impact of Social Features on Video Game Success 3.6.1 Introduction Video games have become very popular: For instance, in the United States video game revenue grew from $10.1 billion in 2009 to $24.5 billion in 2016. Moreover, two-thirds of all US citizens are playing video games on a regular basis (Entertainment Software Association, 2016). However, despite their popularity, the success drivers of video games have not yet been sufficiently empirically investigated. This is an important issue for developers, because the production costs of video games also increased enormously due to new hardware technologies and state-of-the-art game mechanics, graphics, and sound effects. For example, the production budget of “Grand Theft Auto V”, one of the most successful video games of all time, was around $250 million (IMDB, 2016). Therefore, developers can be faced with considerable financial risk nowadays if the video game does not become a market success. To reduce the risk of a low financial results, developers could incorporate paid content through in-app purchases or download content (DLC) so that a user can purchase new weapons, items or new characters and update the basic game. Users who have a high average playtime in the video game are particularly inclined to purchase such additional features (Han et al., 2016). Taking all that into account, developers are facing a huge financial risk when developing a new video game. Nevertheless, taking this risk can certainly pay off. Video games that end up being successful can generate more than 300 percent of their original production costs (Clements and Ohashi, 2005). For example, the previously mentioned video game “Grand Theft Auto V” generated about $800 million in revenue within the first 24 hours of its release (IGN, 2013), as well as over $500 million in the following years through in-app purchases and DLC based on a correspondingly long average playtime (Gamespot, 2016). However, there is high uncertainty among developers regarding which type of video game they should offer, and thus far, success drivers of video games have not been entirely empirically identified. Gretz (2010) showed that the success of a video game depends on a high number of users of a specific video game console, e.g., an Xbox or a PlayStation. Gallaugher and Wang (2002) and Schilling (2002), showed that a high number of users of a specific video game console had a positive effect on the console’s appeal and the sales volume of the available video games. However, the quality of offered video games is an even more important factor regarding the large number of users of a specific video game console (Anderson et al., 2014; Lee, 2013).

3.6 The Need for a Community: The Impact of Social Features on Video Game Success

149

Additional research focused on content-related features of the individual video game. Most authors argued that video games are hedonic products (e.g., Lin, 2010; Turel et al., 2010), the quality of which can only be assessed by a user after the purchase process. Therefore, video games portray economic risks for new users (Voss et al., 2003). To reduce the risk, i.e., the potential failure of the purchase, users frequently rely on reviews, price, the reputation of the developer, and the video game genre. Here, the authors stresses that these features could generate a positive expectation towards the gaming experience and thus lead to a purchase intention. Cox (2013) showed that the reputation of the developer and reviews from journalists could be part of these influential features on the expectation. Moreover, his results show that some genres are getting more preferred than others. Furthermore, Binken and Stremersch (2009) showed a direct positive effect of professional reviews on the expectation and sales of video games. The reviews even had a disproportionate impact when the video game received an absolute top rating of nearly 100 possible points. Nevertheless, the role of newer features, e.g., multiplayer and social interactions with other users while playing the video game, have been ignored as potential success drivers so far. The way video games are consumed by users through social media, online forums, and multiplayer features has changed dramatically within the last ten years (Marchand and Hennig-Thurau, 2013). Therefore, not only the product features impact the expectation and satisfaction towards a video game but also the community respectively social features that has formed around a video game. Through new online multiplayer features, in-app purchases and communication channels, users have been enabled to update the video game, add new features, and share their passion with people from different countries and continents. Video games are no longer just a product that a user buys for a fixed price and enjoys playing on their own or with known friends. Social interaction, knowledge-sharing and lively discussions are essential factors of any online community (Butler, 2001; Wasko and Faraj, 2005) and the video game community thrives on the same principles. Massively multiplayer online games, e.g., “Dota 2” and “Counter-Strike”, enjoy huge popularity among gamers. With an average overall playtime of over 300 hours (lifetime average playtime per gamer), “Counter-Strike” ranks as one of the most played video games on the digital distribution platform Steam (Steam, 2017). Through their participation, users of those video games are not only playing the video game, but are also creators of the gaming experience (BuchananOliver and Seo, 2012). Furthermore, users today even enjoy games in a more passive way. Various YouTube channels that produce primarily “Lets-Play” videos can be seen as a direct result of this development. In this video format, a user films himself while playing a video game and commenting entertainingly on the video game mechanics. At

150

3 Essays

the same time, other users can discuss the video game through chat features or discussion forums (Youtube, 2017). A similar service is accessible through Twitch.tv, a platform that also offers streaming services. Unlike YouTube, this service is more focused on live broadcasting of given material and is especially popular with gaming streams. Once again, users can observe other users playing a video game while interacting through implemented chat functions or even play along with the host through online multiplayer features (Twitch, 2017). Ledbetter and Kuznekoff (2011) argued that the perceived utility of those multiplayer and communication features goes beyond the traditional gaming experience as it enhances interpersonal relationships, and such relationships satisfy a fundamental need for relatedness through interpersonal social interactions within the virtual world (Downie et al., 2008). In summary, video games are increasingly enjoyed as a collaborative social experience and users are keen to discuss and share knowledge about their favorite video games. In summary, previous studies focused only on a limited number of potential success drivers of video games. Thus far, to the best of our knowledge, no study has provided insights into social interactions as success drivers and adequately addressed management implications to enhance and moderate these social interactions. Moreover, as far as we are aware, there are no studies, which explain the actual playtime of a video game. Previous studies mostly focused on the purchase process and the influential factors of a users’ purchase intention, but not on the actual playtime, i.e., the actual use of a video game by users. Mostly the authors argued that the product features directly affect the expectation toward a video game, and therefore the relationship between the product features and sales should be detectable through a market analysis. Therefore, this paper investigates these aspects by transferring the basic assumption of Oliver’s (1977) expectation-confirmation theory (ECT) to this setting. We expand the findings in the field of video games, online communities and information systems by investigating 1) which product features enhance the number of owners of a video game and 2) which product features enhance the average playtime of a given video game. Furthermore, we 3) derive insights into the relative importance of social features in video games in relation to other product features known from the literature. 3.6.2 Conceptual Framework and Hypotheses Development Each video game is differentiated by its graphics, game mechanics, story, sound and multiplayer elements. According to Cooper (1979), the success of such a product innovation depends on the associated value by the users. Unlike pure utilitarian products, video games are hedonic products, as their primary purpose is to create positive emotions (Hirschman and Holbrook, 1982) and joyful and pleasing experiences (Marchand and Hennig-Thurau, 2013). From the perspective of the users, there is high

3.6 The Need for a Community: The Impact of Social Features on Video Game Success

151

uncertainty regarding whether the video game will fulfill this purpose. Since uncertainties are highly subjective and linked to partial information (Garner, 1962), the performance of the video game cannot be entirely predicted due to imperfect information (Salancik and Pfeffer, 1978). Uncertainties in buyer-seller relationships appear mostly due to information asymmetry about the product and the seller (Ghose et al., 2006). In the field of information systems, Dimoka et al. (2012) showed that based on the signal theory, product features, references of the developer and third-party information (such as discussion threads and reviews) signalize a certain value, which results in a specific user expectation of the product respectively video game. Based on the ECT, positive expectations lead to a purchase process and usage of the product. Subsequently, the perceived performance of the game is compared with the user’s expectations. Confirmation results in a satisfying experience, which, according to Oliver (1977), usually leads to repurchase and re-usage intentions. Therefore, based on the ECT, certain video game features, such as the price, genre and references of the developer, reduce the product uncertainty (Dimoka et al., 2012) and form the user expectation regarding whether to buy and play a video game. However, in addition to these primary product features, we extended our framework to take social features into consideration. Today, gamers can communicate, exchange knowledge and support themselves through multiplayer features and chat functions similar to an online community. In general, these social interactions create benefits for each user (Butler et al., 2014) and therefore enhance the overall gaming experience. Users can follow discussions in numerous online forums or streaming services to obtain new information about the video game, e.g., the location of hidden treasures, new items or weapons (Bateman et al., 2011; Kuk, 2006; Millen et al., 2002). For example, there are groups of users who play a video game such as "Minecraft" and connect with each other through online forums and social media channels, which usually leads to motivating them to continue playing the video game (Minecraftforum, 2017). We propose, that these assumptions are in line with network effect theories (e.g., Beck, 2006), which propose that the value for consumers of an membership in a network is positively influenced when new consumers join the network and expand it (Katz and Shapiro, 1994). Consequently, the benefits of consuming such a video game depend positively on the total number of consumers who purchased the latter (Church et al., 2008). Hence, based on previous reasoning, these benefits through social interactions (i.e., network strengthening elements) might become important success drivers for video games (Butler et al., 2014; Wasko and Faraj, 2005), since they can also directly affect the expectations of the video game e.g., based on user reviews and discussion on the product page prior to a purchase, as well as the perceived performance and the intention

152

3 Essays

to play the video game again. Therefore, based on ECT as well as network effects, we analyzed the impact of product features and social features as success drivers of the number of users who bought the video game and the average playtime to gain an understanding of which features attract users to video games and which features motivate them to continue playing. Many studies confirmed that the interest in a product is strongly related to the congruence between the expectation regarding the quality of the product (Ridings and Gefen, 2004; Wasko and Faraj, 2005) and production efforts that meet the current standards. Therefore, it is important to describe the main game features that might lead to expectations due to the product information on the website, i.e., Steam, which in turn could impact the number of owners of the game and the average playtime. Moreover, from a management perspective, these two dependent variables are very important as they shed light on the success drivers of video games, which helps to increase the sales volume of a video game and therefore reduce the production risk. Furthermore, the analysis of the average playtime is an important variable, since a game with a high average playtime usually sells in-apps more frequently (Marchand and Hennig-Thurau, 2013), and thus such features increase the video game revenue. From a research perspective, it is important to gain a better understanding of the impact of social features on hedonic goods, as they may need to be considered more frequently in future research studies. Consequently, we developed the framework shown in Figure 3.6-1. Product Features

Downloadable Content

Developer release history

In-app purchases

Price

Genre & Days since release

Success Features Number of owners of the video game

Social Features

Number of positive votes

Average playtime

Multiplayer

Discussion threads

 Figure 3.6-1. Conceptual Framework

3.6 The Need for a Community: The Impact of Social Features on Video Game Success

153

3.6.3 Impact of the product features Generally, products and software, e.g., video games, differentiate by means of certain product features. Those features create the general conditions in which users decide whether the given game might be worth buying or at least trying out. However, research has shown that there is a process of evaluation regarding these expectations before and after playing the video game. According to Oliver (1977), users create their disconfirmation of beliefs depending on their perceived performance of a product or service and the related expectations of the performance. For instance, with a higher price, the expectations rise and could therefore negatively influence the user behavior due to a high perceived risk that the expectations might not be fulfilled, thus, a high pricing level for a hedonic product increases the perceived chance of a user being disappointed (Voss et al., 2003). Moreover, especially in the context of video games, the rise of “indie games” might even strengthen this challenge. Today, numerous highquality titles are available with only a small monetary cost to users (e.g., “Ori and the Blind Forest”, “Cuphead”). Therefore, the gap between favorable but high-grade titles and full-price video games gets considerably larger so that the latter will be evaluated even more intensively due to the higher pricing. Nonetheless, in the context of video games, research has also shown that the price might also be assessed as part of the quality perception. For instance, Marchand (2016) showed that a higher price has a significant positive influence on the number of owners of video games. Best-selling video games (e.g., “Grand Theft Auto V”) have a very high production budget with state-of-the-art game mechanics, graphics, and sound effects. Therefore, a high price (typically $59.99) increases the interest in the video game and most likely correlates with a high number of owners. However, Marchand (2016) also emphasized that a high price and impressive budgets for state-of-the-art game features mostly correlate with higher budgets for advertisements. Consequently, the number of owners might be affected by aggressive marketing strategies. A similar approach was adopted by Kardes et al. (2004), who showed that there is a certain group of users who always associate a high price with high quality. Nonetheless, this does not apply to every user and might even be interpreted as a special user group (Kardes et al., 2004). These assumptions are in line with behavioral learning theories, which propose that specific pricing triggers expectations that lead to a repeat purchase or not purchase (Rothschild and Gaidis, 1981). Research has shown that price is used as a quality cue, i.e., that price reliance is a tendency of consumers to depend on price as a cue to quality (Zeithaml, 1988). In particular, literature shows a positive response to price when quality is important, i.e., consumers have learned to take price as quality indicator into account and align their behavior accordingly (Tellis and Gaeth, 1990). Especially in this

154

3 Essays

context, most triple-a video games are related to a higher price, shaping the upper boundary of pricing in video games (Marchand, 2016). Hence, we assume that potential users have certain expectations of the video game due to the price, and therefore a video game with a relatively high price might be associated with higher quality regarding the graphics and game mechanics. For this reason, we hypothesize: H1: A higher price of a video game will positively influence the number of owners of a video game. Developers can offer additional content for their games, which provides a new revenue stream as users have to pay for the new content. DLC is additional content created for a released video game which can be downloaded as an upgrade (i.e., update through Steam) by a user. It usually includes new worlds or new characters. Here, DLC mostly distinguishes between directly unlocked content (e.g., the character is instantly selectable as an avatar) or unlockable content (e.g., the character has to be unlocked due to special quests). However, in both cases, the curiosity regarding the new content or the motivation to unlock the content increases the average playtime of a video game. Nonetheless, the content of DLC is noticeably smaller than regular expansions (i.e., a new “arena” vs. completely new games modes), which often leads to a higher number of DLC releases in comparison to regular expansions. For example, the video game “Witcher 3” contains about 19 DLC packages, which combined have a similar playtime to the initially released video game (Steam, 2015). In addition to DLC, the developer can offer in-app purchases, which are usually even smaller packages than DLC. These are small additional items, such as unique virtual weapons or goods, which can provide a direct advantage for a user (Han et al., 2015). However, according to Oliver (1977), those features could also impact the expectations of the users regarding the video game if that additional content is a necessity just to play the video game. If a user is not willing to acquire such content anymore, he may not be able to participate in the virtual world adequately any longer, which makes the video game more and more uninteresting and repetitive (Kane and Alavi, 2007). Such a lack of interest could harm the video game experience for everyone (Farnham et al., 2000). Therefore, the developers should create high-quality DLC and interesting in-app purchases, so that most users are willing to buy such content. If successful, both aspects will tend to increase the average playtime. Therefore, we hypothesize: H2a: In-app features will positively influence the average playtime of a video game. H2b: A large quantity of DLC will positively influence the average playtime of a video game. A successful video game can be expanded into a franchise (i.e., new games or content in the same universe), which is in the interest of a user if he likes the video game.

3.6 The Need for a Community: The Impact of Social Features on Video Game Success

155

However, the realization of sequels and DLC depends heavily on the general success of an established developer. Developers such as Blizzard, Electronic Arts or Ubisoft, with a high number of already published video games, are usually well known to users. Therefore, it is more likely that a game company will produce new sequels and updates in the future as well. Cox (Cox, 2013) showed the significant influence of an established video game developer on the sales of video games. First, the perceived experience (i.e., the game making experience) of users regarding the developer due to published games might be interpreted by users as a quality index for games. Second, a broader history of published games signals to (new) users that more video games will be released in the future, which in turn affects the individual-level decisions (Butler et al., 2014) to buy the video game. Therefore, we transfer the basic assumption of Butler et al. (2014) to the context of video games. Consequently, even though the desire of users to obtain the new desired content only occurs after a satisfying gaming experience, users tend to ensure that this desire can be fulfilled in the first place, i.e., games that could have a follow-up title or content are more favorable. In fact, users are more skeptical about a developer who has only a few titles in the portfolio, as it seems uncertain that such a developer will release sequels and DLC. These assumptions are in line with the associative network theory (Aaker, 1990). Consumers tend to link associations of a brand to form certain expectation towards their products. This, in turn, might create a spillover effect. Therefore, the competence of an established company with a large product portfolio will be transferred to a newly released product (Balachander and Ghose, 2003). Hence, such products appear more attractive to the consumer than a comparable product from an inexperienced or unknown company. For this reason, we also assume that a high number of already published video games signals to a user a certain degree of quality. Consequently, we hypothesize: H3: A high number of published video games by the same developer will positively influence the number of owners of a video game. 3.6.4 Impact of social features In addition to the primary game features in the previous discussion, there might be influential social features that impact video game success which are observable by users before buying the game and might affect potential consumers based on the basic assumption of network theories, that membership in a network (i.e., a specific video game) is positively influenced when new consumers join the network, expand it and enable social interactions (Katz and Shapiro, 1994).

156

3 Essays

For instance, in the context of video games, Cox (2013) showed that reviews are a statistically significant success driver. In principle, reviews have become a comprehensible and straightforward method for users to judge the quality and potential enjoyment of an unknown product. Here, users consider user reviews to be even more trustworthy and a good indicator of the overall quality (Chen and Lurie, 2013; Wu et al., 2013). Therefore, if most users recommend a video game through a positive vote, this reduces the perceived risk and cognitive dissonance in a pre-purchase situation for new users. Chen et al. (2004) confirmed that if the number of consumer votes increases, the overall rating reflects the true quality of a product. Consequently, more user votes could be more persuasive. Regarding the ECT (1977), the number of positive votes will be an indicator of the video game quality (Butler et al., 2014). Based on that idea, we conclude and hypothesize: H4: A higher number of positive votes of a video game will positively influence the number of owners of a video game. Moreover, an online multiplayer feature allows users from different countries to play a video game together. Such a feature is a criterion for social interactions within the video game. Therefore, if a video game has a multiplayer feature, social interaction can usually occur via chat functions or even audio communication (Williams et al., 2007). These technologies allow users to build new social relationships. In addition to the pure video game content (e.g., video game mechanics or graphics), these relationships are perceived as a great benefit by users (Ledbetter and Kuznekoff, 2011). In fact, the participation with each other affects the variety of the users’ individual video game experience (Buchanan-Oliver and Seo, 2012). As previously stated, consumers expanding the video game and thus, enabling a social interaction create benefits for the actual consumer (Church et al., 2008; Katz and Shapiro, 1994). Therefore, each video game tournament and each video game quest will be new and unpredictable and other users directly influence the video game enjoyment. Therefore, we propose that the multiplayer feature is a relevant driver of the average playtime of a video game (Cole and Griffiths, 2007; Hsu and Lu, 2004). The livelier the virtual world gets through the participation of other users, the more diversified the video game experience becomes and the longer and more frequently the users play the video game. Based on this discussion, we hypothesize: H5a: A multiplayer feature will positively influence the number of owners of a video game. H5b: A multiplayer feature will positively influence the average playtime of a video game.

3.6 The Need for a Community: The Impact of Social Features on Video Game Success

157

Based on the findings of Butler et al. (2014), external discussion forums are a further indication of the willingness to participate in the future. A variety of advantages originated in discussion forums, including access to hints (Galegher et al., 1998), support (Ridings and Gefen, 2004), and access to expert knowledge (Lampel and Bhalla, 2007). Such interactions generated by the community lead to a continuous update of the expectations regarding the future gaming experience among users. The intention to continue to participate in the community and thus to continue to play the video game is highly dependent on such expectations (Jin et al., 2010) and creates a perceived benefit (Katz, Haas, et al., 1973). Therefore, we develop the last hypothesis as follows: H6: A large number of discussion threads will positively influence the average playtime of a video game. 3.6.5 Method, Research Design and Sample To test our hypotheses empirically, we collected data on 401 video games from Steam, which has 125 million active users and is the largest online video game distribution platform. The system offers popular video games as well as features for communication and social interactions among the users. Steam is increasingly evolving into a common distribution platform, which therefore enables developers to eliminate the need for a specific publisher. From an empirical point of view, the high quality of the data offers a wide range of analytic approaches. In fact, publishers and developers do not make sales records of video games publicly accessible. For this reason, most of the recent studies (e.g., Cox, 2013; Marchand, 2016) which have examined success drivers of video games have collected data from third-party websites such as vgchartz.com. We argue that if those drivers affect the expectation of users and fulfill them, they should the influence the satisfaction with a videogame, which in turn should be depicted by the predicted relationships through a market analysis on the number of owners and the average playtime. Unfortunately, it is not possible to check the validity of such data. Therefore, we have collected our data directly from the popular platform Steam (Steam, 2018). Table 3.6-1 provides an overview of the operationalization and sources of the variables. Based on our hypotheses, our two dependent variables are the number of owners and the average playtime of the given video game. Steam lists the number of owners who have purchased the given video game on their platform. In our sample, the video game “Counter-Strike” had over 33 million users. The game with the smallest number of users was “Rocket Sky Racing” with only 1,044 users. On average, the video games in our sample had about 1.3 (SD=2.7) million users. Since Steam creates a player profile for each user, it was also possible to determine the average playtime of each video game. In

158

3 Essays

this regard, “Counter-Strike”, with more than 300 hours, was also leading. Conversely, “Smash Pixel Racing” was only played for 10 minutes. In summary, the video games in our sample were played for 25 (SD= 38) hours on average. For each video game, users can recommend the participation through a “thumb up/thumb down” voting system. This system indicates the number of previous users who recommend the video game. Moreover, users can create an endless number of discussion threads for each video game on Steam. Table 3.6-1. Variable operationalization Dependent Variables

Operationalization

Data source

Number of owners of the video game

Number of users who bought the video game.

https://steamdb.info/

Average playtime

Average number of hours the users played the video game.

https://steamdb.info/

Number of positive votes

The number of users who give a positive vote.

https://steamdb.info/

Price

The price at the release of the video game.

https://steamdb.info/

Developer release history

The number of already published games by the same developer.

https://steamdb.info/

Multiplayer

The video game has a multiplayer feature =1, 0 otherwise.

https://steamdb.info/

Discussion threads

The number of discussion threads about the video game.

https://steamcommunity.com/

In-app purchases

The video game has an in-app feature =1, 0 otherwise.

https://steamdb.info/

Downloadble Content (DLC)

The quantity of DLC for the video game

https://steamdb.info/

Days since release

The number of days between the release and the date of the data collection

https://steamdb.info/

Genre

Each genre was set to 1 if appropriate or 0 otherwise

https://steamdb.info/

Independent Variables

Control Variables

We also included this number as an indication of how frequently users discuss the video game. This is a communication function that is heavily used by the users. Some video games have over 100,000 different discussion threads. Furthermore, the price of each video game at its release date was included. Steam also provides information about each developer. With this information, we were able to include the number of previously

3.6 The Need for a Community: The Impact of Social Features on Video Game Success

159

published titles. Moreover, we labeled the multiplayer feature and the in-app purchase feature with a nominal scale (1 and 0) according to whether the video game had such implementation or not. We also included the quantity of available DLC for each video game. As a control variable we included the release date. Thus, we were able to record the period between the release date and the data entry. Because of the rapid decline in the sales volume, in the context of video games, this variable is of particular importance (Cohen, 2014; Moe and Fader, 2001). Nevertheless, we assumed that video games which have been available for a more extended period are more likely to have sold more copies. Therefore, the days since the game was released has been included as a control variable. Consequently, the majority of our sample contains video games which were published between 2007 and 2017. Finally, the genre was also included in our data set as another control variable. In the context of video games, surprisingly little research has been conducted on the feature genre within the discipline of human behavior in information systems. However, comparable research that has been conducted in other fields or industries (e.g., movies and books) provides further evidence that genre has a very important influence on the purchase intention because of personal interests (Austin and Gordan, 1987; Desai and Basuroy, 2005; Elberse and Eliashberg, 2003). Therefore, we included this variable in our calculations as a dummy variable. 3.6.6 Results To test our hypotheses, a structural equation model was computed in SmartPLS. This decision is based on the reasoning that our variables do not represent classical latent constructs, i.e., do not meet the requirement of at least three items, as well as the fact that we checked whether the available data correspond to our assumed model. Here, existing literature recommends the use of the partial least squares method instead of a covariance-based method (Hair et al., 2013). The overall model fit can be evaluated as excellent. Both the NFI (.916) and the SRMR (.044) for our data satisfy the recommended thresholds of NFI >.90 and SRMR